Upload folder using huggingface_hub
Browse filesThis view is limited to 50 files because it contains too many changes.
See raw diff
- 1_Pooling/config.json +10 -0
- README.md +0 -0
- checkpoint-7500/1_Pooling/config.json +10 -0
- checkpoint-7500/README.md +0 -0
- checkpoint-7500/config.json +31 -0
- checkpoint-7500/config_sentence_transformers.json +10 -0
- checkpoint-7500/model.safetensors +3 -0
- checkpoint-7500/modules.json +20 -0
- checkpoint-7500/optimizer.pt +3 -0
- checkpoint-7500/rng_state.pth +3 -0
- checkpoint-7500/scheduler.pt +3 -0
- checkpoint-7500/sentence_bert_config.json +4 -0
- checkpoint-7500/special_tokens_map.json +37 -0
- checkpoint-7500/tokenizer.json +0 -0
- checkpoint-7500/tokenizer_config.json +57 -0
- checkpoint-7500/trainer_state.json +738 -0
- checkpoint-7500/training_args.bin +3 -0
- checkpoint-7500/vocab.txt +0 -0
- checkpoint-8000/1_Pooling/config.json +10 -0
- checkpoint-8000/README.md +0 -0
- checkpoint-8000/config.json +31 -0
- checkpoint-8000/config_sentence_transformers.json +10 -0
- checkpoint-8000/model.safetensors +3 -0
- checkpoint-8000/modules.json +20 -0
- checkpoint-8000/optimizer.pt +3 -0
- checkpoint-8000/rng_state.pth +3 -0
- checkpoint-8000/scheduler.pt +3 -0
- checkpoint-8000/sentence_bert_config.json +4 -0
- checkpoint-8000/special_tokens_map.json +37 -0
- checkpoint-8000/tokenizer.json +0 -0
- checkpoint-8000/tokenizer_config.json +57 -0
- checkpoint-8000/trainer_state.json +785 -0
- checkpoint-8000/training_args.bin +3 -0
- checkpoint-8000/vocab.txt +0 -0
- checkpoint-8500/1_Pooling/config.json +10 -0
- checkpoint-8500/README.md +0 -0
- checkpoint-8500/config.json +31 -0
- checkpoint-8500/config_sentence_transformers.json +10 -0
- checkpoint-8500/model.safetensors +3 -0
- checkpoint-8500/modules.json +20 -0
- checkpoint-8500/optimizer.pt +3 -0
- checkpoint-8500/rng_state.pth +3 -0
- checkpoint-8500/scheduler.pt +3 -0
- checkpoint-8500/sentence_bert_config.json +4 -0
- checkpoint-8500/special_tokens_map.json +37 -0
- checkpoint-8500/tokenizer.json +0 -0
- checkpoint-8500/tokenizer_config.json +57 -0
- checkpoint-8500/trainer_state.json +832 -0
- checkpoint-8500/training_args.bin +3 -0
- checkpoint-8500/vocab.txt +0 -0
1_Pooling/config.json
ADDED
|
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"word_embedding_dimension": 384,
|
| 3 |
+
"pooling_mode_cls_token": true,
|
| 4 |
+
"pooling_mode_mean_tokens": false,
|
| 5 |
+
"pooling_mode_max_tokens": false,
|
| 6 |
+
"pooling_mode_mean_sqrt_len_tokens": false,
|
| 7 |
+
"pooling_mode_weightedmean_tokens": false,
|
| 8 |
+
"pooling_mode_lasttoken": false,
|
| 9 |
+
"include_prompt": true
|
| 10 |
+
}
|
README.md
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
checkpoint-7500/1_Pooling/config.json
ADDED
|
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"word_embedding_dimension": 384,
|
| 3 |
+
"pooling_mode_cls_token": true,
|
| 4 |
+
"pooling_mode_mean_tokens": false,
|
| 5 |
+
"pooling_mode_max_tokens": false,
|
| 6 |
+
"pooling_mode_mean_sqrt_len_tokens": false,
|
| 7 |
+
"pooling_mode_weightedmean_tokens": false,
|
| 8 |
+
"pooling_mode_lasttoken": false,
|
| 9 |
+
"include_prompt": true
|
| 10 |
+
}
|
checkpoint-7500/README.md
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
checkpoint-7500/config.json
ADDED
|
@@ -0,0 +1,31 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "BAAI/bge-small-en-v1.5",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"BertModel"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.1,
|
| 7 |
+
"classifier_dropout": null,
|
| 8 |
+
"hidden_act": "gelu",
|
| 9 |
+
"hidden_dropout_prob": 0.1,
|
| 10 |
+
"hidden_size": 384,
|
| 11 |
+
"id2label": {
|
| 12 |
+
"0": "LABEL_0"
|
| 13 |
+
},
|
| 14 |
+
"initializer_range": 0.02,
|
| 15 |
+
"intermediate_size": 1536,
|
| 16 |
+
"label2id": {
|
| 17 |
+
"LABEL_0": 0
|
| 18 |
+
},
|
| 19 |
+
"layer_norm_eps": 1e-12,
|
| 20 |
+
"max_position_embeddings": 512,
|
| 21 |
+
"model_type": "bert",
|
| 22 |
+
"num_attention_heads": 12,
|
| 23 |
+
"num_hidden_layers": 12,
|
| 24 |
+
"pad_token_id": 0,
|
| 25 |
+
"position_embedding_type": "absolute",
|
| 26 |
+
"torch_dtype": "float32",
|
| 27 |
+
"transformers_version": "4.41.1",
|
| 28 |
+
"type_vocab_size": 2,
|
| 29 |
+
"use_cache": true,
|
| 30 |
+
"vocab_size": 30522
|
| 31 |
+
}
|
checkpoint-7500/config_sentence_transformers.json
ADDED
|
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"__version__": {
|
| 3 |
+
"sentence_transformers": "3.0.1",
|
| 4 |
+
"transformers": "4.41.1",
|
| 5 |
+
"pytorch": "2.3.1+cu121"
|
| 6 |
+
},
|
| 7 |
+
"prompts": {},
|
| 8 |
+
"default_prompt_name": null,
|
| 9 |
+
"similarity_fn_name": null
|
| 10 |
+
}
|
checkpoint-7500/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c4cd0ebce75fee40dfaf85f2f1148c425e21477350fd448712a2dce03a3aa094
|
| 3 |
+
size 133462128
|
checkpoint-7500/modules.json
ADDED
|
@@ -0,0 +1,20 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
[
|
| 2 |
+
{
|
| 3 |
+
"idx": 0,
|
| 4 |
+
"name": "0",
|
| 5 |
+
"path": "",
|
| 6 |
+
"type": "sentence_transformers.models.Transformer"
|
| 7 |
+
},
|
| 8 |
+
{
|
| 9 |
+
"idx": 1,
|
| 10 |
+
"name": "1",
|
| 11 |
+
"path": "1_Pooling",
|
| 12 |
+
"type": "sentence_transformers.models.Pooling"
|
| 13 |
+
},
|
| 14 |
+
{
|
| 15 |
+
"idx": 2,
|
| 16 |
+
"name": "2",
|
| 17 |
+
"path": "2_Normalize",
|
| 18 |
+
"type": "sentence_transformers.models.Normalize"
|
| 19 |
+
}
|
| 20 |
+
]
|
checkpoint-7500/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:24a6efd2207a258804991c9ab74ce8f296de04ad47193b5e6246b7d7391fb5a0
|
| 3 |
+
size 265862074
|
checkpoint-7500/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:81cdeefbf1c010075113334df952d9207ee1e2a63104cb682cff578e50566b33
|
| 3 |
+
size 14244
|
checkpoint-7500/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:88e346cd81efe59235f976f01d8e67ad9f0884cb940e2836af39d12d52ddba59
|
| 3 |
+
size 1064
|
checkpoint-7500/sentence_bert_config.json
ADDED
|
@@ -0,0 +1,4 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"max_seq_length": 512,
|
| 3 |
+
"do_lower_case": true
|
| 4 |
+
}
|
checkpoint-7500/special_tokens_map.json
ADDED
|
@@ -0,0 +1,37 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"cls_token": {
|
| 3 |
+
"content": "[CLS]",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": false,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"mask_token": {
|
| 10 |
+
"content": "[MASK]",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"pad_token": {
|
| 17 |
+
"content": "[PAD]",
|
| 18 |
+
"lstrip": false,
|
| 19 |
+
"normalized": false,
|
| 20 |
+
"rstrip": false,
|
| 21 |
+
"single_word": false
|
| 22 |
+
},
|
| 23 |
+
"sep_token": {
|
| 24 |
+
"content": "[SEP]",
|
| 25 |
+
"lstrip": false,
|
| 26 |
+
"normalized": false,
|
| 27 |
+
"rstrip": false,
|
| 28 |
+
"single_word": false
|
| 29 |
+
},
|
| 30 |
+
"unk_token": {
|
| 31 |
+
"content": "[UNK]",
|
| 32 |
+
"lstrip": false,
|
| 33 |
+
"normalized": false,
|
| 34 |
+
"rstrip": false,
|
| 35 |
+
"single_word": false
|
| 36 |
+
}
|
| 37 |
+
}
|
checkpoint-7500/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
checkpoint-7500/tokenizer_config.json
ADDED
|
@@ -0,0 +1,57 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "[PAD]",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"100": {
|
| 12 |
+
"content": "[UNK]",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"101": {
|
| 20 |
+
"content": "[CLS]",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"102": {
|
| 28 |
+
"content": "[SEP]",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"103": {
|
| 36 |
+
"content": "[MASK]",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"clean_up_tokenization_spaces": true,
|
| 45 |
+
"cls_token": "[CLS]",
|
| 46 |
+
"do_basic_tokenize": true,
|
| 47 |
+
"do_lower_case": true,
|
| 48 |
+
"mask_token": "[MASK]",
|
| 49 |
+
"model_max_length": 512,
|
| 50 |
+
"never_split": null,
|
| 51 |
+
"pad_token": "[PAD]",
|
| 52 |
+
"sep_token": "[SEP]",
|
| 53 |
+
"strip_accents": null,
|
| 54 |
+
"tokenize_chinese_chars": true,
|
| 55 |
+
"tokenizer_class": "BertTokenizer",
|
| 56 |
+
"unk_token": "[UNK]"
|
| 57 |
+
}
|
checkpoint-7500/trainer_state.json
ADDED
|
@@ -0,0 +1,738 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.05420228838920593,
|
| 3 |
+
"best_model_checkpoint": "bge_finetune_dsv2/checkpoint-6500",
|
| 4 |
+
"epoch": 34.09090909090909,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 7500,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 2.2727272727272725,
|
| 13 |
+
"grad_norm": 2.923750162124634,
|
| 14 |
+
"learning_rate": 5.681818181818183e-06,
|
| 15 |
+
"loss": 0.2767,
|
| 16 |
+
"step": 500
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 2.2727272727272725,
|
| 20 |
+
"eval_loss": 0.09305377304553986,
|
| 21 |
+
"eval_runtime": 16.1531,
|
| 22 |
+
"eval_samples_per_second": 48.412,
|
| 23 |
+
"eval_steps_per_second": 6.067,
|
| 24 |
+
"eval_val_evaluator_cosine_accuracy@1": 0.46675191815856776,
|
| 25 |
+
"eval_val_evaluator_cosine_accuracy@10": 0.9411764705882353,
|
| 26 |
+
"eval_val_evaluator_cosine_accuracy@5": 0.8913043478260869,
|
| 27 |
+
"eval_val_evaluator_cosine_map@100": 0.6449144076353992,
|
| 28 |
+
"eval_val_evaluator_cosine_mrr@10": 0.6419812852677294,
|
| 29 |
+
"eval_val_evaluator_cosine_mrr@100": 0.6449144076353982,
|
| 30 |
+
"eval_val_evaluator_cosine_mrr@5": 0.6351449275362308,
|
| 31 |
+
"eval_val_evaluator_cosine_ndcg@10": 0.715922596095415,
|
| 32 |
+
"eval_val_evaluator_cosine_ndcg@100": 0.7286582483109194,
|
| 33 |
+
"eval_val_evaluator_cosine_ndcg@5": 0.6996180454943595,
|
| 34 |
+
"eval_val_evaluator_cosine_precision@1": 0.46675191815856776,
|
| 35 |
+
"eval_val_evaluator_cosine_precision@10": 0.09411764705882353,
|
| 36 |
+
"eval_val_evaluator_cosine_precision@5": 0.17826086956521736,
|
| 37 |
+
"eval_val_evaluator_cosine_recall@1": 0.46675191815856776,
|
| 38 |
+
"eval_val_evaluator_cosine_recall@10": 0.9411764705882353,
|
| 39 |
+
"eval_val_evaluator_cosine_recall@5": 0.8913043478260869,
|
| 40 |
+
"eval_val_evaluator_dot_accuracy@1": 0.46675191815856776,
|
| 41 |
+
"eval_val_evaluator_dot_accuracy@10": 0.9411764705882353,
|
| 42 |
+
"eval_val_evaluator_dot_accuracy@5": 0.8913043478260869,
|
| 43 |
+
"eval_val_evaluator_dot_map@100": 0.6449144076353992,
|
| 44 |
+
"eval_val_evaluator_dot_mrr@10": 0.6419812852677294,
|
| 45 |
+
"eval_val_evaluator_dot_mrr@100": 0.6449144076353982,
|
| 46 |
+
"eval_val_evaluator_dot_mrr@5": 0.6351449275362308,
|
| 47 |
+
"eval_val_evaluator_dot_ndcg@10": 0.715922596095415,
|
| 48 |
+
"eval_val_evaluator_dot_ndcg@100": 0.7286582483109194,
|
| 49 |
+
"eval_val_evaluator_dot_ndcg@5": 0.6996180454943595,
|
| 50 |
+
"eval_val_evaluator_dot_precision@1": 0.46675191815856776,
|
| 51 |
+
"eval_val_evaluator_dot_precision@10": 0.09411764705882353,
|
| 52 |
+
"eval_val_evaluator_dot_precision@5": 0.17826086956521736,
|
| 53 |
+
"eval_val_evaluator_dot_recall@1": 0.46675191815856776,
|
| 54 |
+
"eval_val_evaluator_dot_recall@10": 0.9411764705882353,
|
| 55 |
+
"eval_val_evaluator_dot_recall@5": 0.8913043478260869,
|
| 56 |
+
"step": 500
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 4.545454545454545,
|
| 60 |
+
"grad_norm": 1.0960533618927002,
|
| 61 |
+
"learning_rate": 9.84848484848485e-06,
|
| 62 |
+
"loss": 0.067,
|
| 63 |
+
"step": 1000
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 4.545454545454545,
|
| 67 |
+
"eval_loss": 0.07767628133296967,
|
| 68 |
+
"eval_runtime": 16.162,
|
| 69 |
+
"eval_samples_per_second": 48.385,
|
| 70 |
+
"eval_steps_per_second": 6.064,
|
| 71 |
+
"eval_val_evaluator_cosine_accuracy@1": 0.4680306905370844,
|
| 72 |
+
"eval_val_evaluator_cosine_accuracy@10": 0.948849104859335,
|
| 73 |
+
"eval_val_evaluator_cosine_accuracy@5": 0.8951406649616368,
|
| 74 |
+
"eval_val_evaluator_cosine_map@100": 0.6501350929872222,
|
| 75 |
+
"eval_val_evaluator_cosine_mrr@10": 0.6475830187147313,
|
| 76 |
+
"eval_val_evaluator_cosine_mrr@100": 0.6501350929872212,
|
| 77 |
+
"eval_val_evaluator_cosine_mrr@5": 0.6400468883205445,
|
| 78 |
+
"eval_val_evaluator_cosine_ndcg@10": 0.7221983441504519,
|
| 79 |
+
"eval_val_evaluator_cosine_ndcg@100": 0.7332422183636672,
|
| 80 |
+
"eval_val_evaluator_cosine_ndcg@5": 0.7044596320615103,
|
| 81 |
+
"eval_val_evaluator_cosine_precision@1": 0.4680306905370844,
|
| 82 |
+
"eval_val_evaluator_cosine_precision@10": 0.09488491048593349,
|
| 83 |
+
"eval_val_evaluator_cosine_precision@5": 0.17902813299232737,
|
| 84 |
+
"eval_val_evaluator_cosine_recall@1": 0.4680306905370844,
|
| 85 |
+
"eval_val_evaluator_cosine_recall@10": 0.948849104859335,
|
| 86 |
+
"eval_val_evaluator_cosine_recall@5": 0.8951406649616368,
|
| 87 |
+
"eval_val_evaluator_dot_accuracy@1": 0.4680306905370844,
|
| 88 |
+
"eval_val_evaluator_dot_accuracy@10": 0.948849104859335,
|
| 89 |
+
"eval_val_evaluator_dot_accuracy@5": 0.8951406649616368,
|
| 90 |
+
"eval_val_evaluator_dot_map@100": 0.6501350929872222,
|
| 91 |
+
"eval_val_evaluator_dot_mrr@10": 0.6475830187147313,
|
| 92 |
+
"eval_val_evaluator_dot_mrr@100": 0.6501350929872212,
|
| 93 |
+
"eval_val_evaluator_dot_mrr@5": 0.6400468883205445,
|
| 94 |
+
"eval_val_evaluator_dot_ndcg@10": 0.7221983441504519,
|
| 95 |
+
"eval_val_evaluator_dot_ndcg@100": 0.7332422183636672,
|
| 96 |
+
"eval_val_evaluator_dot_ndcg@5": 0.7044596320615103,
|
| 97 |
+
"eval_val_evaluator_dot_precision@1": 0.4680306905370844,
|
| 98 |
+
"eval_val_evaluator_dot_precision@10": 0.09488491048593349,
|
| 99 |
+
"eval_val_evaluator_dot_precision@5": 0.17902813299232737,
|
| 100 |
+
"eval_val_evaluator_dot_recall@1": 0.4680306905370844,
|
| 101 |
+
"eval_val_evaluator_dot_recall@10": 0.948849104859335,
|
| 102 |
+
"eval_val_evaluator_dot_recall@5": 0.8951406649616368,
|
| 103 |
+
"step": 1000
|
| 104 |
+
},
|
| 105 |
+
{
|
| 106 |
+
"epoch": 6.818181818181818,
|
| 107 |
+
"grad_norm": 1.7721449136734009,
|
| 108 |
+
"learning_rate": 9.217171717171718e-06,
|
| 109 |
+
"loss": 0.0485,
|
| 110 |
+
"step": 1500
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 6.818181818181818,
|
| 114 |
+
"eval_loss": 0.06208512559533119,
|
| 115 |
+
"eval_runtime": 16.1354,
|
| 116 |
+
"eval_samples_per_second": 48.465,
|
| 117 |
+
"eval_steps_per_second": 6.074,
|
| 118 |
+
"eval_val_evaluator_cosine_accuracy@1": 0.4846547314578005,
|
| 119 |
+
"eval_val_evaluator_cosine_accuracy@10": 0.9616368286445013,
|
| 120 |
+
"eval_val_evaluator_cosine_accuracy@5": 0.921994884910486,
|
| 121 |
+
"eval_val_evaluator_cosine_map@100": 0.6677839316557986,
|
| 122 |
+
"eval_val_evaluator_cosine_mrr@10": 0.6659811025859614,
|
| 123 |
+
"eval_val_evaluator_cosine_mrr@100": 0.6677839316557982,
|
| 124 |
+
"eval_val_evaluator_cosine_mrr@5": 0.6605924978687119,
|
| 125 |
+
"eval_val_evaluator_cosine_ndcg@10": 0.739448593920066,
|
| 126 |
+
"eval_val_evaluator_cosine_ndcg@100": 0.7474718746893738,
|
| 127 |
+
"eval_val_evaluator_cosine_ndcg@5": 0.7265352392347777,
|
| 128 |
+
"eval_val_evaluator_cosine_precision@1": 0.4846547314578005,
|
| 129 |
+
"eval_val_evaluator_cosine_precision@10": 0.09616368286445011,
|
| 130 |
+
"eval_val_evaluator_cosine_precision@5": 0.18439897698209717,
|
| 131 |
+
"eval_val_evaluator_cosine_recall@1": 0.4846547314578005,
|
| 132 |
+
"eval_val_evaluator_cosine_recall@10": 0.9616368286445013,
|
| 133 |
+
"eval_val_evaluator_cosine_recall@5": 0.921994884910486,
|
| 134 |
+
"eval_val_evaluator_dot_accuracy@1": 0.4846547314578005,
|
| 135 |
+
"eval_val_evaluator_dot_accuracy@10": 0.9616368286445013,
|
| 136 |
+
"eval_val_evaluator_dot_accuracy@5": 0.921994884910486,
|
| 137 |
+
"eval_val_evaluator_dot_map@100": 0.6677839316557986,
|
| 138 |
+
"eval_val_evaluator_dot_mrr@10": 0.6659811025859614,
|
| 139 |
+
"eval_val_evaluator_dot_mrr@100": 0.6677839316557982,
|
| 140 |
+
"eval_val_evaluator_dot_mrr@5": 0.6605924978687119,
|
| 141 |
+
"eval_val_evaluator_dot_ndcg@10": 0.739448593920066,
|
| 142 |
+
"eval_val_evaluator_dot_ndcg@100": 0.7474718746893738,
|
| 143 |
+
"eval_val_evaluator_dot_ndcg@5": 0.7265352392347777,
|
| 144 |
+
"eval_val_evaluator_dot_precision@1": 0.4846547314578005,
|
| 145 |
+
"eval_val_evaluator_dot_precision@10": 0.09616368286445011,
|
| 146 |
+
"eval_val_evaluator_dot_precision@5": 0.18439897698209717,
|
| 147 |
+
"eval_val_evaluator_dot_recall@1": 0.4846547314578005,
|
| 148 |
+
"eval_val_evaluator_dot_recall@10": 0.9616368286445013,
|
| 149 |
+
"eval_val_evaluator_dot_recall@5": 0.921994884910486,
|
| 150 |
+
"step": 1500
|
| 151 |
+
},
|
| 152 |
+
{
|
| 153 |
+
"epoch": 9.090909090909092,
|
| 154 |
+
"grad_norm": 4.445181846618652,
|
| 155 |
+
"learning_rate": 8.585858585858587e-06,
|
| 156 |
+
"loss": 0.0361,
|
| 157 |
+
"step": 2000
|
| 158 |
+
},
|
| 159 |
+
{
|
| 160 |
+
"epoch": 9.090909090909092,
|
| 161 |
+
"eval_loss": 0.06154213100671768,
|
| 162 |
+
"eval_runtime": 16.1225,
|
| 163 |
+
"eval_samples_per_second": 48.504,
|
| 164 |
+
"eval_steps_per_second": 6.078,
|
| 165 |
+
"eval_val_evaluator_cosine_accuracy@1": 0.4859335038363171,
|
| 166 |
+
"eval_val_evaluator_cosine_accuracy@10": 0.9654731457800512,
|
| 167 |
+
"eval_val_evaluator_cosine_accuracy@5": 0.9271099744245525,
|
| 168 |
+
"eval_val_evaluator_cosine_map@100": 0.6706887408572961,
|
| 169 |
+
"eval_val_evaluator_cosine_mrr@10": 0.6692648073722237,
|
| 170 |
+
"eval_val_evaluator_cosine_mrr@100": 0.6706887408572959,
|
| 171 |
+
"eval_val_evaluator_cosine_mrr@5": 0.6639812446717808,
|
| 172 |
+
"eval_val_evaluator_cosine_ndcg@10": 0.7429450115476419,
|
| 173 |
+
"eval_val_evaluator_cosine_ndcg@100": 0.749838617287201,
|
| 174 |
+
"eval_val_evaluator_cosine_ndcg@5": 0.7303784552891699,
|
| 175 |
+
"eval_val_evaluator_cosine_precision@1": 0.4859335038363171,
|
| 176 |
+
"eval_val_evaluator_cosine_precision@10": 0.09654731457800511,
|
| 177 |
+
"eval_val_evaluator_cosine_precision@5": 0.18542199488491048,
|
| 178 |
+
"eval_val_evaluator_cosine_recall@1": 0.4859335038363171,
|
| 179 |
+
"eval_val_evaluator_cosine_recall@10": 0.9654731457800512,
|
| 180 |
+
"eval_val_evaluator_cosine_recall@5": 0.9271099744245525,
|
| 181 |
+
"eval_val_evaluator_dot_accuracy@1": 0.4859335038363171,
|
| 182 |
+
"eval_val_evaluator_dot_accuracy@10": 0.9654731457800512,
|
| 183 |
+
"eval_val_evaluator_dot_accuracy@5": 0.9271099744245525,
|
| 184 |
+
"eval_val_evaluator_dot_map@100": 0.6706887408572961,
|
| 185 |
+
"eval_val_evaluator_dot_mrr@10": 0.6692648073722237,
|
| 186 |
+
"eval_val_evaluator_dot_mrr@100": 0.6706887408572959,
|
| 187 |
+
"eval_val_evaluator_dot_mrr@5": 0.6639812446717808,
|
| 188 |
+
"eval_val_evaluator_dot_ndcg@10": 0.7429450115476419,
|
| 189 |
+
"eval_val_evaluator_dot_ndcg@100": 0.749838617287201,
|
| 190 |
+
"eval_val_evaluator_dot_ndcg@5": 0.7303784552891699,
|
| 191 |
+
"eval_val_evaluator_dot_precision@1": 0.4859335038363171,
|
| 192 |
+
"eval_val_evaluator_dot_precision@10": 0.09654731457800511,
|
| 193 |
+
"eval_val_evaluator_dot_precision@5": 0.18542199488491048,
|
| 194 |
+
"eval_val_evaluator_dot_recall@1": 0.4859335038363171,
|
| 195 |
+
"eval_val_evaluator_dot_recall@10": 0.9654731457800512,
|
| 196 |
+
"eval_val_evaluator_dot_recall@5": 0.9271099744245525,
|
| 197 |
+
"step": 2000
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 11.363636363636363,
|
| 201 |
+
"grad_norm": 7.720225811004639,
|
| 202 |
+
"learning_rate": 7.954545454545455e-06,
|
| 203 |
+
"loss": 0.0301,
|
| 204 |
+
"step": 2500
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 11.363636363636363,
|
| 208 |
+
"eval_loss": 0.06867850571870804,
|
| 209 |
+
"eval_runtime": 16.1559,
|
| 210 |
+
"eval_samples_per_second": 48.403,
|
| 211 |
+
"eval_steps_per_second": 6.066,
|
| 212 |
+
"eval_val_evaluator_cosine_accuracy@1": 0.49232736572890023,
|
| 213 |
+
"eval_val_evaluator_cosine_accuracy@10": 0.9603580562659847,
|
| 214 |
+
"eval_val_evaluator_cosine_accuracy@5": 0.928388746803069,
|
| 215 |
+
"eval_val_evaluator_cosine_map@100": 0.6764517431036522,
|
| 216 |
+
"eval_val_evaluator_cosine_mrr@10": 0.6745762797872769,
|
| 217 |
+
"eval_val_evaluator_cosine_mrr@100": 0.6764517431036515,
|
| 218 |
+
"eval_val_evaluator_cosine_mrr@5": 0.6698635976129574,
|
| 219 |
+
"eval_val_evaluator_cosine_ndcg@10": 0.7459669546464727,
|
| 220 |
+
"eval_val_evaluator_cosine_ndcg@100": 0.7541025758659824,
|
| 221 |
+
"eval_val_evaluator_cosine_ndcg@5": 0.7351837372969686,
|
| 222 |
+
"eval_val_evaluator_cosine_precision@1": 0.49232736572890023,
|
| 223 |
+
"eval_val_evaluator_cosine_precision@10": 0.09603580562659846,
|
| 224 |
+
"eval_val_evaluator_cosine_precision@5": 0.1856777493606138,
|
| 225 |
+
"eval_val_evaluator_cosine_recall@1": 0.49232736572890023,
|
| 226 |
+
"eval_val_evaluator_cosine_recall@10": 0.9603580562659847,
|
| 227 |
+
"eval_val_evaluator_cosine_recall@5": 0.928388746803069,
|
| 228 |
+
"eval_val_evaluator_dot_accuracy@1": 0.49232736572890023,
|
| 229 |
+
"eval_val_evaluator_dot_accuracy@10": 0.9603580562659847,
|
| 230 |
+
"eval_val_evaluator_dot_accuracy@5": 0.928388746803069,
|
| 231 |
+
"eval_val_evaluator_dot_map@100": 0.6764517431036522,
|
| 232 |
+
"eval_val_evaluator_dot_mrr@10": 0.6745762797872769,
|
| 233 |
+
"eval_val_evaluator_dot_mrr@100": 0.6764517431036515,
|
| 234 |
+
"eval_val_evaluator_dot_mrr@5": 0.6698635976129574,
|
| 235 |
+
"eval_val_evaluator_dot_ndcg@10": 0.7459669546464727,
|
| 236 |
+
"eval_val_evaluator_dot_ndcg@100": 0.7541025758659824,
|
| 237 |
+
"eval_val_evaluator_dot_ndcg@5": 0.7351837372969686,
|
| 238 |
+
"eval_val_evaluator_dot_precision@1": 0.49232736572890023,
|
| 239 |
+
"eval_val_evaluator_dot_precision@10": 0.09603580562659846,
|
| 240 |
+
"eval_val_evaluator_dot_precision@5": 0.1856777493606138,
|
| 241 |
+
"eval_val_evaluator_dot_recall@1": 0.49232736572890023,
|
| 242 |
+
"eval_val_evaluator_dot_recall@10": 0.9603580562659847,
|
| 243 |
+
"eval_val_evaluator_dot_recall@5": 0.928388746803069,
|
| 244 |
+
"step": 2500
|
| 245 |
+
},
|
| 246 |
+
{
|
| 247 |
+
"epoch": 13.636363636363637,
|
| 248 |
+
"grad_norm": 3.5265557765960693,
|
| 249 |
+
"learning_rate": 7.323232323232324e-06,
|
| 250 |
+
"loss": 0.0274,
|
| 251 |
+
"step": 3000
|
| 252 |
+
},
|
| 253 |
+
{
|
| 254 |
+
"epoch": 13.636363636363637,
|
| 255 |
+
"eval_loss": 0.06610127538442612,
|
| 256 |
+
"eval_runtime": 16.1447,
|
| 257 |
+
"eval_samples_per_second": 48.437,
|
| 258 |
+
"eval_steps_per_second": 6.07,
|
| 259 |
+
"eval_val_evaluator_cosine_accuracy@1": 0.48721227621483376,
|
| 260 |
+
"eval_val_evaluator_cosine_accuracy@10": 0.9603580562659847,
|
| 261 |
+
"eval_val_evaluator_cosine_accuracy@5": 0.9245524296675192,
|
| 262 |
+
"eval_val_evaluator_cosine_map@100": 0.6732921649074852,
|
| 263 |
+
"eval_val_evaluator_cosine_mrr@10": 0.671490987699427,
|
| 264 |
+
"eval_val_evaluator_cosine_mrr@100": 0.6732921649074848,
|
| 265 |
+
"eval_val_evaluator_cosine_mrr@5": 0.6664748508098882,
|
| 266 |
+
"eval_val_evaluator_cosine_ndcg@10": 0.7435664719550824,
|
| 267 |
+
"eval_val_evaluator_cosine_ndcg@100": 0.7519530408382193,
|
| 268 |
+
"eval_val_evaluator_cosine_ndcg@5": 0.7317557792940828,
|
| 269 |
+
"eval_val_evaluator_cosine_precision@1": 0.48721227621483376,
|
| 270 |
+
"eval_val_evaluator_cosine_precision@10": 0.09603580562659846,
|
| 271 |
+
"eval_val_evaluator_cosine_precision@5": 0.18491048593350382,
|
| 272 |
+
"eval_val_evaluator_cosine_recall@1": 0.48721227621483376,
|
| 273 |
+
"eval_val_evaluator_cosine_recall@10": 0.9603580562659847,
|
| 274 |
+
"eval_val_evaluator_cosine_recall@5": 0.9245524296675192,
|
| 275 |
+
"eval_val_evaluator_dot_accuracy@1": 0.48721227621483376,
|
| 276 |
+
"eval_val_evaluator_dot_accuracy@10": 0.9603580562659847,
|
| 277 |
+
"eval_val_evaluator_dot_accuracy@5": 0.9245524296675192,
|
| 278 |
+
"eval_val_evaluator_dot_map@100": 0.6732921649074852,
|
| 279 |
+
"eval_val_evaluator_dot_mrr@10": 0.671490987699427,
|
| 280 |
+
"eval_val_evaluator_dot_mrr@100": 0.6732921649074848,
|
| 281 |
+
"eval_val_evaluator_dot_mrr@5": 0.6664748508098882,
|
| 282 |
+
"eval_val_evaluator_dot_ndcg@10": 0.7435664719550824,
|
| 283 |
+
"eval_val_evaluator_dot_ndcg@100": 0.7519530408382193,
|
| 284 |
+
"eval_val_evaluator_dot_ndcg@5": 0.7317557792940828,
|
| 285 |
+
"eval_val_evaluator_dot_precision@1": 0.48721227621483376,
|
| 286 |
+
"eval_val_evaluator_dot_precision@10": 0.09603580562659846,
|
| 287 |
+
"eval_val_evaluator_dot_precision@5": 0.18491048593350382,
|
| 288 |
+
"eval_val_evaluator_dot_recall@1": 0.48721227621483376,
|
| 289 |
+
"eval_val_evaluator_dot_recall@10": 0.9603580562659847,
|
| 290 |
+
"eval_val_evaluator_dot_recall@5": 0.9245524296675192,
|
| 291 |
+
"step": 3000
|
| 292 |
+
},
|
| 293 |
+
{
|
| 294 |
+
"epoch": 15.909090909090908,
|
| 295 |
+
"grad_norm": 2.537912130355835,
|
| 296 |
+
"learning_rate": 6.691919191919193e-06,
|
| 297 |
+
"loss": 0.0223,
|
| 298 |
+
"step": 3500
|
| 299 |
+
},
|
| 300 |
+
{
|
| 301 |
+
"epoch": 15.909090909090908,
|
| 302 |
+
"eval_loss": 0.06057285517454147,
|
| 303 |
+
"eval_runtime": 16.2937,
|
| 304 |
+
"eval_samples_per_second": 47.994,
|
| 305 |
+
"eval_steps_per_second": 6.015,
|
| 306 |
+
"eval_val_evaluator_cosine_accuracy@1": 0.5051150895140665,
|
| 307 |
+
"eval_val_evaluator_cosine_accuracy@10": 0.959079283887468,
|
| 308 |
+
"eval_val_evaluator_cosine_accuracy@5": 0.928388746803069,
|
| 309 |
+
"eval_val_evaluator_cosine_map@100": 0.6822101368749947,
|
| 310 |
+
"eval_val_evaluator_cosine_mrr@10": 0.680220131530873,
|
| 311 |
+
"eval_val_evaluator_cosine_mrr@100": 0.6822101368749943,
|
| 312 |
+
"eval_val_evaluator_cosine_mrr@5": 0.6759164535379364,
|
| 313 |
+
"eval_val_evaluator_cosine_ndcg@10": 0.749745056809794,
|
| 314 |
+
"eval_val_evaluator_cosine_ndcg@100": 0.7586065473558236,
|
| 315 |
+
"eval_val_evaluator_cosine_ndcg@5": 0.7396139266559898,
|
| 316 |
+
"eval_val_evaluator_cosine_precision@1": 0.5051150895140665,
|
| 317 |
+
"eval_val_evaluator_cosine_precision@10": 0.0959079283887468,
|
| 318 |
+
"eval_val_evaluator_cosine_precision@5": 0.1856777493606138,
|
| 319 |
+
"eval_val_evaluator_cosine_recall@1": 0.5051150895140665,
|
| 320 |
+
"eval_val_evaluator_cosine_recall@10": 0.959079283887468,
|
| 321 |
+
"eval_val_evaluator_cosine_recall@5": 0.928388746803069,
|
| 322 |
+
"eval_val_evaluator_dot_accuracy@1": 0.5051150895140665,
|
| 323 |
+
"eval_val_evaluator_dot_accuracy@10": 0.959079283887468,
|
| 324 |
+
"eval_val_evaluator_dot_accuracy@5": 0.928388746803069,
|
| 325 |
+
"eval_val_evaluator_dot_map@100": 0.6822101368749947,
|
| 326 |
+
"eval_val_evaluator_dot_mrr@10": 0.680220131530873,
|
| 327 |
+
"eval_val_evaluator_dot_mrr@100": 0.6822101368749943,
|
| 328 |
+
"eval_val_evaluator_dot_mrr@5": 0.6759164535379364,
|
| 329 |
+
"eval_val_evaluator_dot_ndcg@10": 0.749745056809794,
|
| 330 |
+
"eval_val_evaluator_dot_ndcg@100": 0.7586065473558236,
|
| 331 |
+
"eval_val_evaluator_dot_ndcg@5": 0.7396139266559898,
|
| 332 |
+
"eval_val_evaluator_dot_precision@1": 0.5051150895140665,
|
| 333 |
+
"eval_val_evaluator_dot_precision@10": 0.0959079283887468,
|
| 334 |
+
"eval_val_evaluator_dot_precision@5": 0.1856777493606138,
|
| 335 |
+
"eval_val_evaluator_dot_recall@1": 0.5051150895140665,
|
| 336 |
+
"eval_val_evaluator_dot_recall@10": 0.959079283887468,
|
| 337 |
+
"eval_val_evaluator_dot_recall@5": 0.928388746803069,
|
| 338 |
+
"step": 3500
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 18.181818181818183,
|
| 342 |
+
"grad_norm": 6.390940189361572,
|
| 343 |
+
"learning_rate": 6.060606060606061e-06,
|
| 344 |
+
"loss": 0.021,
|
| 345 |
+
"step": 4000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 18.181818181818183,
|
| 349 |
+
"eval_loss": 0.05628308653831482,
|
| 350 |
+
"eval_runtime": 16.1509,
|
| 351 |
+
"eval_samples_per_second": 48.418,
|
| 352 |
+
"eval_steps_per_second": 6.068,
|
| 353 |
+
"eval_val_evaluator_cosine_accuracy@1": 0.5063938618925832,
|
| 354 |
+
"eval_val_evaluator_cosine_accuracy@10": 0.9552429667519181,
|
| 355 |
+
"eval_val_evaluator_cosine_accuracy@5": 0.9245524296675192,
|
| 356 |
+
"eval_val_evaluator_cosine_map@100": 0.6833838272532422,
|
| 357 |
+
"eval_val_evaluator_cosine_mrr@10": 0.6811883449031781,
|
| 358 |
+
"eval_val_evaluator_cosine_mrr@100": 0.6833838272532414,
|
| 359 |
+
"eval_val_evaluator_cosine_mrr@5": 0.6767689684569473,
|
| 360 |
+
"eval_val_evaluator_cosine_ndcg@10": 0.7496397009089436,
|
| 361 |
+
"eval_val_evaluator_cosine_ndcg@100": 0.7591798763800799,
|
| 362 |
+
"eval_val_evaluator_cosine_ndcg@5": 0.7393934605063354,
|
| 363 |
+
"eval_val_evaluator_cosine_precision@1": 0.5063938618925832,
|
| 364 |
+
"eval_val_evaluator_cosine_precision@10": 0.0955242966751918,
|
| 365 |
+
"eval_val_evaluator_cosine_precision@5": 0.1849104859335038,
|
| 366 |
+
"eval_val_evaluator_cosine_recall@1": 0.5063938618925832,
|
| 367 |
+
"eval_val_evaluator_cosine_recall@10": 0.9552429667519181,
|
| 368 |
+
"eval_val_evaluator_cosine_recall@5": 0.9245524296675192,
|
| 369 |
+
"eval_val_evaluator_dot_accuracy@1": 0.5063938618925832,
|
| 370 |
+
"eval_val_evaluator_dot_accuracy@10": 0.9552429667519181,
|
| 371 |
+
"eval_val_evaluator_dot_accuracy@5": 0.9245524296675192,
|
| 372 |
+
"eval_val_evaluator_dot_map@100": 0.6833838272532422,
|
| 373 |
+
"eval_val_evaluator_dot_mrr@10": 0.6811883449031781,
|
| 374 |
+
"eval_val_evaluator_dot_mrr@100": 0.6833838272532414,
|
| 375 |
+
"eval_val_evaluator_dot_mrr@5": 0.6767689684569473,
|
| 376 |
+
"eval_val_evaluator_dot_ndcg@10": 0.7496397009089436,
|
| 377 |
+
"eval_val_evaluator_dot_ndcg@100": 0.7591798763800799,
|
| 378 |
+
"eval_val_evaluator_dot_ndcg@5": 0.7393934605063354,
|
| 379 |
+
"eval_val_evaluator_dot_precision@1": 0.5063938618925832,
|
| 380 |
+
"eval_val_evaluator_dot_precision@10": 0.0955242966751918,
|
| 381 |
+
"eval_val_evaluator_dot_precision@5": 0.1849104859335038,
|
| 382 |
+
"eval_val_evaluator_dot_recall@1": 0.5063938618925832,
|
| 383 |
+
"eval_val_evaluator_dot_recall@10": 0.9552429667519181,
|
| 384 |
+
"eval_val_evaluator_dot_recall@5": 0.9245524296675192,
|
| 385 |
+
"step": 4000
|
| 386 |
+
},
|
| 387 |
+
{
|
| 388 |
+
"epoch": 20.454545454545453,
|
| 389 |
+
"grad_norm": 4.392763614654541,
|
| 390 |
+
"learning_rate": 5.429292929292929e-06,
|
| 391 |
+
"loss": 0.0203,
|
| 392 |
+
"step": 4500
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 20.454545454545453,
|
| 396 |
+
"eval_loss": 0.057314660400152206,
|
| 397 |
+
"eval_runtime": 16.1911,
|
| 398 |
+
"eval_samples_per_second": 48.298,
|
| 399 |
+
"eval_steps_per_second": 6.053,
|
| 400 |
+
"eval_val_evaluator_cosine_accuracy@1": 0.4782608695652174,
|
| 401 |
+
"eval_val_evaluator_cosine_accuracy@10": 0.9578005115089514,
|
| 402 |
+
"eval_val_evaluator_cosine_accuracy@5": 0.928388746803069,
|
| 403 |
+
"eval_val_evaluator_cosine_map@100": 0.6680531566193644,
|
| 404 |
+
"eval_val_evaluator_cosine_mrr@10": 0.6661059351276735,
|
| 405 |
+
"eval_val_evaluator_cosine_mrr@100": 0.6680531566193635,
|
| 406 |
+
"eval_val_evaluator_cosine_mrr@5": 0.6618925831202036,
|
| 407 |
+
"eval_val_evaluator_cosine_ndcg@10": 0.7390313134229293,
|
| 408 |
+
"eval_val_evaluator_cosine_ndcg@100": 0.7478157002187452,
|
| 409 |
+
"eval_val_evaluator_cosine_ndcg@5": 0.7292375411745403,
|
| 410 |
+
"eval_val_evaluator_cosine_precision@1": 0.4782608695652174,
|
| 411 |
+
"eval_val_evaluator_cosine_precision@10": 0.09578005115089515,
|
| 412 |
+
"eval_val_evaluator_cosine_precision@5": 0.1856777493606138,
|
| 413 |
+
"eval_val_evaluator_cosine_recall@1": 0.4782608695652174,
|
| 414 |
+
"eval_val_evaluator_cosine_recall@10": 0.9578005115089514,
|
| 415 |
+
"eval_val_evaluator_cosine_recall@5": 0.928388746803069,
|
| 416 |
+
"eval_val_evaluator_dot_accuracy@1": 0.4782608695652174,
|
| 417 |
+
"eval_val_evaluator_dot_accuracy@10": 0.9578005115089514,
|
| 418 |
+
"eval_val_evaluator_dot_accuracy@5": 0.928388746803069,
|
| 419 |
+
"eval_val_evaluator_dot_map@100": 0.6680531566193644,
|
| 420 |
+
"eval_val_evaluator_dot_mrr@10": 0.6661059351276735,
|
| 421 |
+
"eval_val_evaluator_dot_mrr@100": 0.6680531566193635,
|
| 422 |
+
"eval_val_evaluator_dot_mrr@5": 0.6618925831202036,
|
| 423 |
+
"eval_val_evaluator_dot_ndcg@10": 0.7390313134229293,
|
| 424 |
+
"eval_val_evaluator_dot_ndcg@100": 0.7478157002187452,
|
| 425 |
+
"eval_val_evaluator_dot_ndcg@5": 0.7292375411745403,
|
| 426 |
+
"eval_val_evaluator_dot_precision@1": 0.4782608695652174,
|
| 427 |
+
"eval_val_evaluator_dot_precision@10": 0.09578005115089515,
|
| 428 |
+
"eval_val_evaluator_dot_precision@5": 0.1856777493606138,
|
| 429 |
+
"eval_val_evaluator_dot_recall@1": 0.4782608695652174,
|
| 430 |
+
"eval_val_evaluator_dot_recall@10": 0.9578005115089514,
|
| 431 |
+
"eval_val_evaluator_dot_recall@5": 0.928388746803069,
|
| 432 |
+
"step": 4500
|
| 433 |
+
},
|
| 434 |
+
{
|
| 435 |
+
"epoch": 22.727272727272727,
|
| 436 |
+
"grad_norm": 4.935870170593262,
|
| 437 |
+
"learning_rate": 4.7979797979797985e-06,
|
| 438 |
+
"loss": 0.0212,
|
| 439 |
+
"step": 5000
|
| 440 |
+
},
|
| 441 |
+
{
|
| 442 |
+
"epoch": 22.727272727272727,
|
| 443 |
+
"eval_loss": 0.06370092183351517,
|
| 444 |
+
"eval_runtime": 16.1455,
|
| 445 |
+
"eval_samples_per_second": 48.435,
|
| 446 |
+
"eval_steps_per_second": 6.07,
|
| 447 |
+
"eval_val_evaluator_cosine_accuracy@1": 0.49872122762148335,
|
| 448 |
+
"eval_val_evaluator_cosine_accuracy@10": 0.9514066496163683,
|
| 449 |
+
"eval_val_evaluator_cosine_accuracy@5": 0.9258312020460358,
|
| 450 |
+
"eval_val_evaluator_cosine_map@100": 0.6769910008307545,
|
| 451 |
+
"eval_val_evaluator_cosine_mrr@10": 0.6745772946859899,
|
| 452 |
+
"eval_val_evaluator_cosine_mrr@100": 0.6769910008307539,
|
| 453 |
+
"eval_val_evaluator_cosine_mrr@5": 0.6707587382779191,
|
| 454 |
+
"eval_val_evaluator_cosine_ndcg@10": 0.743803897516351,
|
| 455 |
+
"eval_val_evaluator_cosine_ndcg@100": 0.7540671552876991,
|
| 456 |
+
"eval_val_evaluator_cosine_ndcg@5": 0.7351253519366493,
|
| 457 |
+
"eval_val_evaluator_cosine_precision@1": 0.49872122762148335,
|
| 458 |
+
"eval_val_evaluator_cosine_precision@10": 0.0951406649616368,
|
| 459 |
+
"eval_val_evaluator_cosine_precision@5": 0.18516624040920715,
|
| 460 |
+
"eval_val_evaluator_cosine_recall@1": 0.49872122762148335,
|
| 461 |
+
"eval_val_evaluator_cosine_recall@10": 0.9514066496163683,
|
| 462 |
+
"eval_val_evaluator_cosine_recall@5": 0.9258312020460358,
|
| 463 |
+
"eval_val_evaluator_dot_accuracy@1": 0.49872122762148335,
|
| 464 |
+
"eval_val_evaluator_dot_accuracy@10": 0.9514066496163683,
|
| 465 |
+
"eval_val_evaluator_dot_accuracy@5": 0.9258312020460358,
|
| 466 |
+
"eval_val_evaluator_dot_map@100": 0.6769910008307545,
|
| 467 |
+
"eval_val_evaluator_dot_mrr@10": 0.6745772946859899,
|
| 468 |
+
"eval_val_evaluator_dot_mrr@100": 0.6769910008307539,
|
| 469 |
+
"eval_val_evaluator_dot_mrr@5": 0.6707587382779191,
|
| 470 |
+
"eval_val_evaluator_dot_ndcg@10": 0.743803897516351,
|
| 471 |
+
"eval_val_evaluator_dot_ndcg@100": 0.7540671552876991,
|
| 472 |
+
"eval_val_evaluator_dot_ndcg@5": 0.7351253519366493,
|
| 473 |
+
"eval_val_evaluator_dot_precision@1": 0.49872122762148335,
|
| 474 |
+
"eval_val_evaluator_dot_precision@10": 0.0951406649616368,
|
| 475 |
+
"eval_val_evaluator_dot_precision@5": 0.18516624040920715,
|
| 476 |
+
"eval_val_evaluator_dot_recall@1": 0.49872122762148335,
|
| 477 |
+
"eval_val_evaluator_dot_recall@10": 0.9514066496163683,
|
| 478 |
+
"eval_val_evaluator_dot_recall@5": 0.9258312020460358,
|
| 479 |
+
"step": 5000
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 25.0,
|
| 483 |
+
"grad_norm": 8.184098243713379,
|
| 484 |
+
"learning_rate": 4.166666666666667e-06,
|
| 485 |
+
"loss": 0.018,
|
| 486 |
+
"step": 5500
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 25.0,
|
| 490 |
+
"eval_loss": 0.057994794100522995,
|
| 491 |
+
"eval_runtime": 16.1282,
|
| 492 |
+
"eval_samples_per_second": 48.487,
|
| 493 |
+
"eval_steps_per_second": 6.076,
|
| 494 |
+
"eval_val_evaluator_cosine_accuracy@1": 0.49744245524296676,
|
| 495 |
+
"eval_val_evaluator_cosine_accuracy@10": 0.9539641943734015,
|
| 496 |
+
"eval_val_evaluator_cosine_accuracy@5": 0.9245524296675192,
|
| 497 |
+
"eval_val_evaluator_cosine_map@100": 0.6780504881130015,
|
| 498 |
+
"eval_val_evaluator_cosine_mrr@10": 0.6757890837494416,
|
| 499 |
+
"eval_val_evaluator_cosine_mrr@100": 0.6780504881130013,
|
| 500 |
+
"eval_val_evaluator_cosine_mrr@5": 0.6714407502131281,
|
| 501 |
+
"eval_val_evaluator_cosine_ndcg@10": 0.7453518958660976,
|
| 502 |
+
"eval_val_evaluator_cosine_ndcg@100": 0.7550072905243007,
|
| 503 |
+
"eval_val_evaluator_cosine_ndcg@5": 0.7354200452387065,
|
| 504 |
+
"eval_val_evaluator_cosine_precision@1": 0.49744245524296676,
|
| 505 |
+
"eval_val_evaluator_cosine_precision@10": 0.09539641943734015,
|
| 506 |
+
"eval_val_evaluator_cosine_precision@5": 0.18491048593350382,
|
| 507 |
+
"eval_val_evaluator_cosine_recall@1": 0.49744245524296676,
|
| 508 |
+
"eval_val_evaluator_cosine_recall@10": 0.9539641943734015,
|
| 509 |
+
"eval_val_evaluator_cosine_recall@5": 0.9245524296675192,
|
| 510 |
+
"eval_val_evaluator_dot_accuracy@1": 0.49744245524296676,
|
| 511 |
+
"eval_val_evaluator_dot_accuracy@10": 0.9539641943734015,
|
| 512 |
+
"eval_val_evaluator_dot_accuracy@5": 0.9245524296675192,
|
| 513 |
+
"eval_val_evaluator_dot_map@100": 0.6780504881130015,
|
| 514 |
+
"eval_val_evaluator_dot_mrr@10": 0.6757890837494416,
|
| 515 |
+
"eval_val_evaluator_dot_mrr@100": 0.6780504881130013,
|
| 516 |
+
"eval_val_evaluator_dot_mrr@5": 0.6714407502131281,
|
| 517 |
+
"eval_val_evaluator_dot_ndcg@10": 0.7453518958660976,
|
| 518 |
+
"eval_val_evaluator_dot_ndcg@100": 0.7550072905243007,
|
| 519 |
+
"eval_val_evaluator_dot_ndcg@5": 0.7354200452387065,
|
| 520 |
+
"eval_val_evaluator_dot_precision@1": 0.49744245524296676,
|
| 521 |
+
"eval_val_evaluator_dot_precision@10": 0.09539641943734015,
|
| 522 |
+
"eval_val_evaluator_dot_precision@5": 0.18491048593350382,
|
| 523 |
+
"eval_val_evaluator_dot_recall@1": 0.49744245524296676,
|
| 524 |
+
"eval_val_evaluator_dot_recall@10": 0.9539641943734015,
|
| 525 |
+
"eval_val_evaluator_dot_recall@5": 0.9245524296675192,
|
| 526 |
+
"step": 5500
|
| 527 |
+
},
|
| 528 |
+
{
|
| 529 |
+
"epoch": 27.272727272727273,
|
| 530 |
+
"grad_norm": 0.004099330864846706,
|
| 531 |
+
"learning_rate": 3.5353535353535356e-06,
|
| 532 |
+
"loss": 0.0166,
|
| 533 |
+
"step": 6000
|
| 534 |
+
},
|
| 535 |
+
{
|
| 536 |
+
"epoch": 27.272727272727273,
|
| 537 |
+
"eval_loss": 0.056713227182626724,
|
| 538 |
+
"eval_runtime": 16.1679,
|
| 539 |
+
"eval_samples_per_second": 48.367,
|
| 540 |
+
"eval_steps_per_second": 6.061,
|
| 541 |
+
"eval_val_evaluator_cosine_accuracy@1": 0.4948849104859335,
|
| 542 |
+
"eval_val_evaluator_cosine_accuracy@10": 0.9552429667519181,
|
| 543 |
+
"eval_val_evaluator_cosine_accuracy@5": 0.9232736572890026,
|
| 544 |
+
"eval_val_evaluator_cosine_map@100": 0.6780550124960388,
|
| 545 |
+
"eval_val_evaluator_cosine_mrr@10": 0.675847440425445,
|
| 546 |
+
"eval_val_evaluator_cosine_mrr@100": 0.6780550124960384,
|
| 547 |
+
"eval_val_evaluator_cosine_mrr@5": 0.6710358056265976,
|
| 548 |
+
"eval_val_evaluator_cosine_ndcg@10": 0.7457753760186268,
|
| 549 |
+
"eval_val_evaluator_cosine_ndcg@100": 0.7551473066673557,
|
| 550 |
+
"eval_val_evaluator_cosine_ndcg@5": 0.7348943591506862,
|
| 551 |
+
"eval_val_evaluator_cosine_precision@1": 0.4948849104859335,
|
| 552 |
+
"eval_val_evaluator_cosine_precision@10": 0.0955242966751918,
|
| 553 |
+
"eval_val_evaluator_cosine_precision@5": 0.18465473145780048,
|
| 554 |
+
"eval_val_evaluator_cosine_recall@1": 0.4948849104859335,
|
| 555 |
+
"eval_val_evaluator_cosine_recall@10": 0.9552429667519181,
|
| 556 |
+
"eval_val_evaluator_cosine_recall@5": 0.9232736572890026,
|
| 557 |
+
"eval_val_evaluator_dot_accuracy@1": 0.4948849104859335,
|
| 558 |
+
"eval_val_evaluator_dot_accuracy@10": 0.9552429667519181,
|
| 559 |
+
"eval_val_evaluator_dot_accuracy@5": 0.9232736572890026,
|
| 560 |
+
"eval_val_evaluator_dot_map@100": 0.6780550124960388,
|
| 561 |
+
"eval_val_evaluator_dot_mrr@10": 0.675847440425445,
|
| 562 |
+
"eval_val_evaluator_dot_mrr@100": 0.6780550124960384,
|
| 563 |
+
"eval_val_evaluator_dot_mrr@5": 0.6710358056265976,
|
| 564 |
+
"eval_val_evaluator_dot_ndcg@10": 0.7457753760186268,
|
| 565 |
+
"eval_val_evaluator_dot_ndcg@100": 0.7551473066673557,
|
| 566 |
+
"eval_val_evaluator_dot_ndcg@5": 0.7348943591506862,
|
| 567 |
+
"eval_val_evaluator_dot_precision@1": 0.4948849104859335,
|
| 568 |
+
"eval_val_evaluator_dot_precision@10": 0.0955242966751918,
|
| 569 |
+
"eval_val_evaluator_dot_precision@5": 0.18465473145780048,
|
| 570 |
+
"eval_val_evaluator_dot_recall@1": 0.4948849104859335,
|
| 571 |
+
"eval_val_evaluator_dot_recall@10": 0.9552429667519181,
|
| 572 |
+
"eval_val_evaluator_dot_recall@5": 0.9232736572890026,
|
| 573 |
+
"step": 6000
|
| 574 |
+
},
|
| 575 |
+
{
|
| 576 |
+
"epoch": 29.545454545454547,
|
| 577 |
+
"grad_norm": 0.02236221544444561,
|
| 578 |
+
"learning_rate": 2.904040404040404e-06,
|
| 579 |
+
"loss": 0.0194,
|
| 580 |
+
"step": 6500
|
| 581 |
+
},
|
| 582 |
+
{
|
| 583 |
+
"epoch": 29.545454545454547,
|
| 584 |
+
"eval_loss": 0.05420228838920593,
|
| 585 |
+
"eval_runtime": 16.1682,
|
| 586 |
+
"eval_samples_per_second": 48.366,
|
| 587 |
+
"eval_steps_per_second": 6.061,
|
| 588 |
+
"eval_val_evaluator_cosine_accuracy@1": 0.5038363171355499,
|
| 589 |
+
"eval_val_evaluator_cosine_accuracy@10": 0.9565217391304348,
|
| 590 |
+
"eval_val_evaluator_cosine_accuracy@5": 0.9271099744245525,
|
| 591 |
+
"eval_val_evaluator_cosine_map@100": 0.6834975753216492,
|
| 592 |
+
"eval_val_evaluator_cosine_mrr@10": 0.6813415946088577,
|
| 593 |
+
"eval_val_evaluator_cosine_mrr@100": 0.6834975753216488,
|
| 594 |
+
"eval_val_evaluator_cosine_mrr@5": 0.6769181585677743,
|
| 595 |
+
"eval_val_evaluator_cosine_ndcg@10": 0.7501703504734094,
|
| 596 |
+
"eval_val_evaluator_cosine_ndcg@100": 0.759281214123215,
|
| 597 |
+
"eval_val_evaluator_cosine_ndcg@5": 0.7401646982131597,
|
| 598 |
+
"eval_val_evaluator_cosine_precision@1": 0.5038363171355499,
|
| 599 |
+
"eval_val_evaluator_cosine_precision@10": 0.09565217391304348,
|
| 600 |
+
"eval_val_evaluator_cosine_precision@5": 0.18542199488491046,
|
| 601 |
+
"eval_val_evaluator_cosine_recall@1": 0.5038363171355499,
|
| 602 |
+
"eval_val_evaluator_cosine_recall@10": 0.9565217391304348,
|
| 603 |
+
"eval_val_evaluator_cosine_recall@5": 0.9271099744245525,
|
| 604 |
+
"eval_val_evaluator_dot_accuracy@1": 0.5038363171355499,
|
| 605 |
+
"eval_val_evaluator_dot_accuracy@10": 0.9565217391304348,
|
| 606 |
+
"eval_val_evaluator_dot_accuracy@5": 0.9271099744245525,
|
| 607 |
+
"eval_val_evaluator_dot_map@100": 0.6834975753216492,
|
| 608 |
+
"eval_val_evaluator_dot_mrr@10": 0.6813415946088577,
|
| 609 |
+
"eval_val_evaluator_dot_mrr@100": 0.6834975753216488,
|
| 610 |
+
"eval_val_evaluator_dot_mrr@5": 0.6769181585677743,
|
| 611 |
+
"eval_val_evaluator_dot_ndcg@10": 0.7501703504734094,
|
| 612 |
+
"eval_val_evaluator_dot_ndcg@100": 0.759281214123215,
|
| 613 |
+
"eval_val_evaluator_dot_ndcg@5": 0.7401646982131597,
|
| 614 |
+
"eval_val_evaluator_dot_precision@1": 0.5038363171355499,
|
| 615 |
+
"eval_val_evaluator_dot_precision@10": 0.09565217391304348,
|
| 616 |
+
"eval_val_evaluator_dot_precision@5": 0.18542199488491046,
|
| 617 |
+
"eval_val_evaluator_dot_recall@1": 0.5038363171355499,
|
| 618 |
+
"eval_val_evaluator_dot_recall@10": 0.9565217391304348,
|
| 619 |
+
"eval_val_evaluator_dot_recall@5": 0.9271099744245525,
|
| 620 |
+
"step": 6500
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 31.818181818181817,
|
| 624 |
+
"grad_norm": 0.03244137763977051,
|
| 625 |
+
"learning_rate": 2.2727272727272728e-06,
|
| 626 |
+
"loss": 0.0182,
|
| 627 |
+
"step": 7000
|
| 628 |
+
},
|
| 629 |
+
{
|
| 630 |
+
"epoch": 31.818181818181817,
|
| 631 |
+
"eval_loss": 0.05468086898326874,
|
| 632 |
+
"eval_runtime": 23.5877,
|
| 633 |
+
"eval_samples_per_second": 33.153,
|
| 634 |
+
"eval_steps_per_second": 4.155,
|
| 635 |
+
"eval_val_evaluator_cosine_accuracy@1": 0.5140664961636828,
|
| 636 |
+
"eval_val_evaluator_cosine_accuracy@10": 0.9578005115089514,
|
| 637 |
+
"eval_val_evaluator_cosine_accuracy@5": 0.9296675191815856,
|
| 638 |
+
"eval_val_evaluator_cosine_map@100": 0.6896503713996865,
|
| 639 |
+
"eval_val_evaluator_cosine_mrr@10": 0.6875806844476918,
|
| 640 |
+
"eval_val_evaluator_cosine_mrr@100": 0.6896503713996863,
|
| 641 |
+
"eval_val_evaluator_cosine_mrr@5": 0.6833333333333328,
|
| 642 |
+
"eval_val_evaluator_cosine_ndcg@10": 0.7551953164170838,
|
| 643 |
+
"eval_val_evaluator_cosine_ndcg@100": 0.7640059344945531,
|
| 644 |
+
"eval_val_evaluator_cosine_ndcg@5": 0.7456070216519929,
|
| 645 |
+
"eval_val_evaluator_cosine_precision@1": 0.5140664961636828,
|
| 646 |
+
"eval_val_evaluator_cosine_precision@10": 0.09578005115089515,
|
| 647 |
+
"eval_val_evaluator_cosine_precision@5": 0.1859335038363171,
|
| 648 |
+
"eval_val_evaluator_cosine_recall@1": 0.5140664961636828,
|
| 649 |
+
"eval_val_evaluator_cosine_recall@10": 0.9578005115089514,
|
| 650 |
+
"eval_val_evaluator_cosine_recall@5": 0.9296675191815856,
|
| 651 |
+
"eval_val_evaluator_dot_accuracy@1": 0.5140664961636828,
|
| 652 |
+
"eval_val_evaluator_dot_accuracy@10": 0.9578005115089514,
|
| 653 |
+
"eval_val_evaluator_dot_accuracy@5": 0.9296675191815856,
|
| 654 |
+
"eval_val_evaluator_dot_map@100": 0.6896503713996865,
|
| 655 |
+
"eval_val_evaluator_dot_mrr@10": 0.6875806844476918,
|
| 656 |
+
"eval_val_evaluator_dot_mrr@100": 0.6896503713996863,
|
| 657 |
+
"eval_val_evaluator_dot_mrr@5": 0.6833333333333328,
|
| 658 |
+
"eval_val_evaluator_dot_ndcg@10": 0.7551953164170838,
|
| 659 |
+
"eval_val_evaluator_dot_ndcg@100": 0.7640059344945531,
|
| 660 |
+
"eval_val_evaluator_dot_ndcg@5": 0.7456070216519929,
|
| 661 |
+
"eval_val_evaluator_dot_precision@1": 0.5140664961636828,
|
| 662 |
+
"eval_val_evaluator_dot_precision@10": 0.09578005115089515,
|
| 663 |
+
"eval_val_evaluator_dot_precision@5": 0.1859335038363171,
|
| 664 |
+
"eval_val_evaluator_dot_recall@1": 0.5140664961636828,
|
| 665 |
+
"eval_val_evaluator_dot_recall@10": 0.9578005115089514,
|
| 666 |
+
"eval_val_evaluator_dot_recall@5": 0.9296675191815856,
|
| 667 |
+
"step": 7000
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 34.09090909090909,
|
| 671 |
+
"grad_norm": 0.001623653806746006,
|
| 672 |
+
"learning_rate": 1.6414141414141415e-06,
|
| 673 |
+
"loss": 0.0157,
|
| 674 |
+
"step": 7500
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 34.09090909090909,
|
| 678 |
+
"eval_loss": 0.05486880987882614,
|
| 679 |
+
"eval_runtime": 16.1005,
|
| 680 |
+
"eval_samples_per_second": 48.57,
|
| 681 |
+
"eval_steps_per_second": 6.087,
|
| 682 |
+
"eval_val_evaluator_cosine_accuracy@1": 0.5140664961636828,
|
| 683 |
+
"eval_val_evaluator_cosine_accuracy@10": 0.9565217391304348,
|
| 684 |
+
"eval_val_evaluator_cosine_accuracy@5": 0.9309462915601023,
|
| 685 |
+
"eval_val_evaluator_cosine_map@100": 0.6898639676357106,
|
| 686 |
+
"eval_val_evaluator_cosine_mrr@10": 0.6877618438679816,
|
| 687 |
+
"eval_val_evaluator_cosine_mrr@100": 0.6898639676357101,
|
| 688 |
+
"eval_val_evaluator_cosine_mrr@5": 0.6839727195225911,
|
| 689 |
+
"eval_val_evaluator_cosine_ndcg@10": 0.7550036263650965,
|
| 690 |
+
"eval_val_evaluator_cosine_ndcg@100": 0.764061133822081,
|
| 691 |
+
"eval_val_evaluator_cosine_ndcg@5": 0.7463601070791411,
|
| 692 |
+
"eval_val_evaluator_cosine_precision@1": 0.5140664961636828,
|
| 693 |
+
"eval_val_evaluator_cosine_precision@10": 0.09565217391304345,
|
| 694 |
+
"eval_val_evaluator_cosine_precision@5": 0.18618925831202046,
|
| 695 |
+
"eval_val_evaluator_cosine_recall@1": 0.5140664961636828,
|
| 696 |
+
"eval_val_evaluator_cosine_recall@10": 0.9565217391304348,
|
| 697 |
+
"eval_val_evaluator_cosine_recall@5": 0.9309462915601023,
|
| 698 |
+
"eval_val_evaluator_dot_accuracy@1": 0.5140664961636828,
|
| 699 |
+
"eval_val_evaluator_dot_accuracy@10": 0.9565217391304348,
|
| 700 |
+
"eval_val_evaluator_dot_accuracy@5": 0.9309462915601023,
|
| 701 |
+
"eval_val_evaluator_dot_map@100": 0.6898639676357106,
|
| 702 |
+
"eval_val_evaluator_dot_mrr@10": 0.6877618438679816,
|
| 703 |
+
"eval_val_evaluator_dot_mrr@100": 0.6898639676357101,
|
| 704 |
+
"eval_val_evaluator_dot_mrr@5": 0.6839727195225911,
|
| 705 |
+
"eval_val_evaluator_dot_ndcg@10": 0.7550036263650965,
|
| 706 |
+
"eval_val_evaluator_dot_ndcg@100": 0.764061133822081,
|
| 707 |
+
"eval_val_evaluator_dot_ndcg@5": 0.7463601070791411,
|
| 708 |
+
"eval_val_evaluator_dot_precision@1": 0.5140664961636828,
|
| 709 |
+
"eval_val_evaluator_dot_precision@10": 0.09565217391304345,
|
| 710 |
+
"eval_val_evaluator_dot_precision@5": 0.18618925831202046,
|
| 711 |
+
"eval_val_evaluator_dot_recall@1": 0.5140664961636828,
|
| 712 |
+
"eval_val_evaluator_dot_recall@10": 0.9565217391304348,
|
| 713 |
+
"eval_val_evaluator_dot_recall@5": 0.9309462915601023,
|
| 714 |
+
"step": 7500
|
| 715 |
+
}
|
| 716 |
+
],
|
| 717 |
+
"logging_steps": 500,
|
| 718 |
+
"max_steps": 8800,
|
| 719 |
+
"num_input_tokens_seen": 0,
|
| 720 |
+
"num_train_epochs": 40,
|
| 721 |
+
"save_steps": 500,
|
| 722 |
+
"stateful_callbacks": {
|
| 723 |
+
"TrainerControl": {
|
| 724 |
+
"args": {
|
| 725 |
+
"should_epoch_stop": false,
|
| 726 |
+
"should_evaluate": false,
|
| 727 |
+
"should_log": false,
|
| 728 |
+
"should_save": true,
|
| 729 |
+
"should_training_stop": false
|
| 730 |
+
},
|
| 731 |
+
"attributes": {}
|
| 732 |
+
}
|
| 733 |
+
},
|
| 734 |
+
"total_flos": 0.0,
|
| 735 |
+
"train_batch_size": 8,
|
| 736 |
+
"trial_name": null,
|
| 737 |
+
"trial_params": null
|
| 738 |
+
}
|
checkpoint-7500/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:47648917f49529e1cd5c4f55390eb54738fe83ae8b04b361ce299a8b142ad8a8
|
| 3 |
+
size 5304
|
checkpoint-7500/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
checkpoint-8000/1_Pooling/config.json
ADDED
|
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"word_embedding_dimension": 384,
|
| 3 |
+
"pooling_mode_cls_token": true,
|
| 4 |
+
"pooling_mode_mean_tokens": false,
|
| 5 |
+
"pooling_mode_max_tokens": false,
|
| 6 |
+
"pooling_mode_mean_sqrt_len_tokens": false,
|
| 7 |
+
"pooling_mode_weightedmean_tokens": false,
|
| 8 |
+
"pooling_mode_lasttoken": false,
|
| 9 |
+
"include_prompt": true
|
| 10 |
+
}
|
checkpoint-8000/README.md
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
checkpoint-8000/config.json
ADDED
|
@@ -0,0 +1,31 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "BAAI/bge-small-en-v1.5",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"BertModel"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.1,
|
| 7 |
+
"classifier_dropout": null,
|
| 8 |
+
"hidden_act": "gelu",
|
| 9 |
+
"hidden_dropout_prob": 0.1,
|
| 10 |
+
"hidden_size": 384,
|
| 11 |
+
"id2label": {
|
| 12 |
+
"0": "LABEL_0"
|
| 13 |
+
},
|
| 14 |
+
"initializer_range": 0.02,
|
| 15 |
+
"intermediate_size": 1536,
|
| 16 |
+
"label2id": {
|
| 17 |
+
"LABEL_0": 0
|
| 18 |
+
},
|
| 19 |
+
"layer_norm_eps": 1e-12,
|
| 20 |
+
"max_position_embeddings": 512,
|
| 21 |
+
"model_type": "bert",
|
| 22 |
+
"num_attention_heads": 12,
|
| 23 |
+
"num_hidden_layers": 12,
|
| 24 |
+
"pad_token_id": 0,
|
| 25 |
+
"position_embedding_type": "absolute",
|
| 26 |
+
"torch_dtype": "float32",
|
| 27 |
+
"transformers_version": "4.41.1",
|
| 28 |
+
"type_vocab_size": 2,
|
| 29 |
+
"use_cache": true,
|
| 30 |
+
"vocab_size": 30522
|
| 31 |
+
}
|
checkpoint-8000/config_sentence_transformers.json
ADDED
|
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"__version__": {
|
| 3 |
+
"sentence_transformers": "3.0.1",
|
| 4 |
+
"transformers": "4.41.1",
|
| 5 |
+
"pytorch": "2.3.1+cu121"
|
| 6 |
+
},
|
| 7 |
+
"prompts": {},
|
| 8 |
+
"default_prompt_name": null,
|
| 9 |
+
"similarity_fn_name": null
|
| 10 |
+
}
|
checkpoint-8000/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:99b0721b41552acc69333bb8eabbcb6c761127670d79f31c962be3b936b2f2d7
|
| 3 |
+
size 133462128
|
checkpoint-8000/modules.json
ADDED
|
@@ -0,0 +1,20 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
[
|
| 2 |
+
{
|
| 3 |
+
"idx": 0,
|
| 4 |
+
"name": "0",
|
| 5 |
+
"path": "",
|
| 6 |
+
"type": "sentence_transformers.models.Transformer"
|
| 7 |
+
},
|
| 8 |
+
{
|
| 9 |
+
"idx": 1,
|
| 10 |
+
"name": "1",
|
| 11 |
+
"path": "1_Pooling",
|
| 12 |
+
"type": "sentence_transformers.models.Pooling"
|
| 13 |
+
},
|
| 14 |
+
{
|
| 15 |
+
"idx": 2,
|
| 16 |
+
"name": "2",
|
| 17 |
+
"path": "2_Normalize",
|
| 18 |
+
"type": "sentence_transformers.models.Normalize"
|
| 19 |
+
}
|
| 20 |
+
]
|
checkpoint-8000/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:eda486f030a722ec5ff6689a59ec79fcf0f1f00e06eed5b60cd7cd9568bcec27
|
| 3 |
+
size 265862074
|
checkpoint-8000/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:72442dec6e39b771e40a1bf3032fd15ba340bc468e666d6229119212928e42f7
|
| 3 |
+
size 14244
|
checkpoint-8000/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c101e1f4ea8e2ddf3842b06a4d6779f2076390889dd0290f3d7bf8f4f5909129
|
| 3 |
+
size 1064
|
checkpoint-8000/sentence_bert_config.json
ADDED
|
@@ -0,0 +1,4 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"max_seq_length": 512,
|
| 3 |
+
"do_lower_case": true
|
| 4 |
+
}
|
checkpoint-8000/special_tokens_map.json
ADDED
|
@@ -0,0 +1,37 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"cls_token": {
|
| 3 |
+
"content": "[CLS]",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": false,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"mask_token": {
|
| 10 |
+
"content": "[MASK]",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"pad_token": {
|
| 17 |
+
"content": "[PAD]",
|
| 18 |
+
"lstrip": false,
|
| 19 |
+
"normalized": false,
|
| 20 |
+
"rstrip": false,
|
| 21 |
+
"single_word": false
|
| 22 |
+
},
|
| 23 |
+
"sep_token": {
|
| 24 |
+
"content": "[SEP]",
|
| 25 |
+
"lstrip": false,
|
| 26 |
+
"normalized": false,
|
| 27 |
+
"rstrip": false,
|
| 28 |
+
"single_word": false
|
| 29 |
+
},
|
| 30 |
+
"unk_token": {
|
| 31 |
+
"content": "[UNK]",
|
| 32 |
+
"lstrip": false,
|
| 33 |
+
"normalized": false,
|
| 34 |
+
"rstrip": false,
|
| 35 |
+
"single_word": false
|
| 36 |
+
}
|
| 37 |
+
}
|
checkpoint-8000/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
checkpoint-8000/tokenizer_config.json
ADDED
|
@@ -0,0 +1,57 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "[PAD]",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"100": {
|
| 12 |
+
"content": "[UNK]",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"101": {
|
| 20 |
+
"content": "[CLS]",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"102": {
|
| 28 |
+
"content": "[SEP]",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"103": {
|
| 36 |
+
"content": "[MASK]",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"clean_up_tokenization_spaces": true,
|
| 45 |
+
"cls_token": "[CLS]",
|
| 46 |
+
"do_basic_tokenize": true,
|
| 47 |
+
"do_lower_case": true,
|
| 48 |
+
"mask_token": "[MASK]",
|
| 49 |
+
"model_max_length": 512,
|
| 50 |
+
"never_split": null,
|
| 51 |
+
"pad_token": "[PAD]",
|
| 52 |
+
"sep_token": "[SEP]",
|
| 53 |
+
"strip_accents": null,
|
| 54 |
+
"tokenize_chinese_chars": true,
|
| 55 |
+
"tokenizer_class": "BertTokenizer",
|
| 56 |
+
"unk_token": "[UNK]"
|
| 57 |
+
}
|
checkpoint-8000/trainer_state.json
ADDED
|
@@ -0,0 +1,785 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.05301735922694206,
|
| 3 |
+
"best_model_checkpoint": "bge_finetune_dsv2/checkpoint-8000",
|
| 4 |
+
"epoch": 36.36363636363637,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 8000,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 2.2727272727272725,
|
| 13 |
+
"grad_norm": 2.923750162124634,
|
| 14 |
+
"learning_rate": 5.681818181818183e-06,
|
| 15 |
+
"loss": 0.2767,
|
| 16 |
+
"step": 500
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 2.2727272727272725,
|
| 20 |
+
"eval_loss": 0.09305377304553986,
|
| 21 |
+
"eval_runtime": 16.1531,
|
| 22 |
+
"eval_samples_per_second": 48.412,
|
| 23 |
+
"eval_steps_per_second": 6.067,
|
| 24 |
+
"eval_val_evaluator_cosine_accuracy@1": 0.46675191815856776,
|
| 25 |
+
"eval_val_evaluator_cosine_accuracy@10": 0.9411764705882353,
|
| 26 |
+
"eval_val_evaluator_cosine_accuracy@5": 0.8913043478260869,
|
| 27 |
+
"eval_val_evaluator_cosine_map@100": 0.6449144076353992,
|
| 28 |
+
"eval_val_evaluator_cosine_mrr@10": 0.6419812852677294,
|
| 29 |
+
"eval_val_evaluator_cosine_mrr@100": 0.6449144076353982,
|
| 30 |
+
"eval_val_evaluator_cosine_mrr@5": 0.6351449275362308,
|
| 31 |
+
"eval_val_evaluator_cosine_ndcg@10": 0.715922596095415,
|
| 32 |
+
"eval_val_evaluator_cosine_ndcg@100": 0.7286582483109194,
|
| 33 |
+
"eval_val_evaluator_cosine_ndcg@5": 0.6996180454943595,
|
| 34 |
+
"eval_val_evaluator_cosine_precision@1": 0.46675191815856776,
|
| 35 |
+
"eval_val_evaluator_cosine_precision@10": 0.09411764705882353,
|
| 36 |
+
"eval_val_evaluator_cosine_precision@5": 0.17826086956521736,
|
| 37 |
+
"eval_val_evaluator_cosine_recall@1": 0.46675191815856776,
|
| 38 |
+
"eval_val_evaluator_cosine_recall@10": 0.9411764705882353,
|
| 39 |
+
"eval_val_evaluator_cosine_recall@5": 0.8913043478260869,
|
| 40 |
+
"eval_val_evaluator_dot_accuracy@1": 0.46675191815856776,
|
| 41 |
+
"eval_val_evaluator_dot_accuracy@10": 0.9411764705882353,
|
| 42 |
+
"eval_val_evaluator_dot_accuracy@5": 0.8913043478260869,
|
| 43 |
+
"eval_val_evaluator_dot_map@100": 0.6449144076353992,
|
| 44 |
+
"eval_val_evaluator_dot_mrr@10": 0.6419812852677294,
|
| 45 |
+
"eval_val_evaluator_dot_mrr@100": 0.6449144076353982,
|
| 46 |
+
"eval_val_evaluator_dot_mrr@5": 0.6351449275362308,
|
| 47 |
+
"eval_val_evaluator_dot_ndcg@10": 0.715922596095415,
|
| 48 |
+
"eval_val_evaluator_dot_ndcg@100": 0.7286582483109194,
|
| 49 |
+
"eval_val_evaluator_dot_ndcg@5": 0.6996180454943595,
|
| 50 |
+
"eval_val_evaluator_dot_precision@1": 0.46675191815856776,
|
| 51 |
+
"eval_val_evaluator_dot_precision@10": 0.09411764705882353,
|
| 52 |
+
"eval_val_evaluator_dot_precision@5": 0.17826086956521736,
|
| 53 |
+
"eval_val_evaluator_dot_recall@1": 0.46675191815856776,
|
| 54 |
+
"eval_val_evaluator_dot_recall@10": 0.9411764705882353,
|
| 55 |
+
"eval_val_evaluator_dot_recall@5": 0.8913043478260869,
|
| 56 |
+
"step": 500
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 4.545454545454545,
|
| 60 |
+
"grad_norm": 1.0960533618927002,
|
| 61 |
+
"learning_rate": 9.84848484848485e-06,
|
| 62 |
+
"loss": 0.067,
|
| 63 |
+
"step": 1000
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 4.545454545454545,
|
| 67 |
+
"eval_loss": 0.07767628133296967,
|
| 68 |
+
"eval_runtime": 16.162,
|
| 69 |
+
"eval_samples_per_second": 48.385,
|
| 70 |
+
"eval_steps_per_second": 6.064,
|
| 71 |
+
"eval_val_evaluator_cosine_accuracy@1": 0.4680306905370844,
|
| 72 |
+
"eval_val_evaluator_cosine_accuracy@10": 0.948849104859335,
|
| 73 |
+
"eval_val_evaluator_cosine_accuracy@5": 0.8951406649616368,
|
| 74 |
+
"eval_val_evaluator_cosine_map@100": 0.6501350929872222,
|
| 75 |
+
"eval_val_evaluator_cosine_mrr@10": 0.6475830187147313,
|
| 76 |
+
"eval_val_evaluator_cosine_mrr@100": 0.6501350929872212,
|
| 77 |
+
"eval_val_evaluator_cosine_mrr@5": 0.6400468883205445,
|
| 78 |
+
"eval_val_evaluator_cosine_ndcg@10": 0.7221983441504519,
|
| 79 |
+
"eval_val_evaluator_cosine_ndcg@100": 0.7332422183636672,
|
| 80 |
+
"eval_val_evaluator_cosine_ndcg@5": 0.7044596320615103,
|
| 81 |
+
"eval_val_evaluator_cosine_precision@1": 0.4680306905370844,
|
| 82 |
+
"eval_val_evaluator_cosine_precision@10": 0.09488491048593349,
|
| 83 |
+
"eval_val_evaluator_cosine_precision@5": 0.17902813299232737,
|
| 84 |
+
"eval_val_evaluator_cosine_recall@1": 0.4680306905370844,
|
| 85 |
+
"eval_val_evaluator_cosine_recall@10": 0.948849104859335,
|
| 86 |
+
"eval_val_evaluator_cosine_recall@5": 0.8951406649616368,
|
| 87 |
+
"eval_val_evaluator_dot_accuracy@1": 0.4680306905370844,
|
| 88 |
+
"eval_val_evaluator_dot_accuracy@10": 0.948849104859335,
|
| 89 |
+
"eval_val_evaluator_dot_accuracy@5": 0.8951406649616368,
|
| 90 |
+
"eval_val_evaluator_dot_map@100": 0.6501350929872222,
|
| 91 |
+
"eval_val_evaluator_dot_mrr@10": 0.6475830187147313,
|
| 92 |
+
"eval_val_evaluator_dot_mrr@100": 0.6501350929872212,
|
| 93 |
+
"eval_val_evaluator_dot_mrr@5": 0.6400468883205445,
|
| 94 |
+
"eval_val_evaluator_dot_ndcg@10": 0.7221983441504519,
|
| 95 |
+
"eval_val_evaluator_dot_ndcg@100": 0.7332422183636672,
|
| 96 |
+
"eval_val_evaluator_dot_ndcg@5": 0.7044596320615103,
|
| 97 |
+
"eval_val_evaluator_dot_precision@1": 0.4680306905370844,
|
| 98 |
+
"eval_val_evaluator_dot_precision@10": 0.09488491048593349,
|
| 99 |
+
"eval_val_evaluator_dot_precision@5": 0.17902813299232737,
|
| 100 |
+
"eval_val_evaluator_dot_recall@1": 0.4680306905370844,
|
| 101 |
+
"eval_val_evaluator_dot_recall@10": 0.948849104859335,
|
| 102 |
+
"eval_val_evaluator_dot_recall@5": 0.8951406649616368,
|
| 103 |
+
"step": 1000
|
| 104 |
+
},
|
| 105 |
+
{
|
| 106 |
+
"epoch": 6.818181818181818,
|
| 107 |
+
"grad_norm": 1.7721449136734009,
|
| 108 |
+
"learning_rate": 9.217171717171718e-06,
|
| 109 |
+
"loss": 0.0485,
|
| 110 |
+
"step": 1500
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 6.818181818181818,
|
| 114 |
+
"eval_loss": 0.06208512559533119,
|
| 115 |
+
"eval_runtime": 16.1354,
|
| 116 |
+
"eval_samples_per_second": 48.465,
|
| 117 |
+
"eval_steps_per_second": 6.074,
|
| 118 |
+
"eval_val_evaluator_cosine_accuracy@1": 0.4846547314578005,
|
| 119 |
+
"eval_val_evaluator_cosine_accuracy@10": 0.9616368286445013,
|
| 120 |
+
"eval_val_evaluator_cosine_accuracy@5": 0.921994884910486,
|
| 121 |
+
"eval_val_evaluator_cosine_map@100": 0.6677839316557986,
|
| 122 |
+
"eval_val_evaluator_cosine_mrr@10": 0.6659811025859614,
|
| 123 |
+
"eval_val_evaluator_cosine_mrr@100": 0.6677839316557982,
|
| 124 |
+
"eval_val_evaluator_cosine_mrr@5": 0.6605924978687119,
|
| 125 |
+
"eval_val_evaluator_cosine_ndcg@10": 0.739448593920066,
|
| 126 |
+
"eval_val_evaluator_cosine_ndcg@100": 0.7474718746893738,
|
| 127 |
+
"eval_val_evaluator_cosine_ndcg@5": 0.7265352392347777,
|
| 128 |
+
"eval_val_evaluator_cosine_precision@1": 0.4846547314578005,
|
| 129 |
+
"eval_val_evaluator_cosine_precision@10": 0.09616368286445011,
|
| 130 |
+
"eval_val_evaluator_cosine_precision@5": 0.18439897698209717,
|
| 131 |
+
"eval_val_evaluator_cosine_recall@1": 0.4846547314578005,
|
| 132 |
+
"eval_val_evaluator_cosine_recall@10": 0.9616368286445013,
|
| 133 |
+
"eval_val_evaluator_cosine_recall@5": 0.921994884910486,
|
| 134 |
+
"eval_val_evaluator_dot_accuracy@1": 0.4846547314578005,
|
| 135 |
+
"eval_val_evaluator_dot_accuracy@10": 0.9616368286445013,
|
| 136 |
+
"eval_val_evaluator_dot_accuracy@5": 0.921994884910486,
|
| 137 |
+
"eval_val_evaluator_dot_map@100": 0.6677839316557986,
|
| 138 |
+
"eval_val_evaluator_dot_mrr@10": 0.6659811025859614,
|
| 139 |
+
"eval_val_evaluator_dot_mrr@100": 0.6677839316557982,
|
| 140 |
+
"eval_val_evaluator_dot_mrr@5": 0.6605924978687119,
|
| 141 |
+
"eval_val_evaluator_dot_ndcg@10": 0.739448593920066,
|
| 142 |
+
"eval_val_evaluator_dot_ndcg@100": 0.7474718746893738,
|
| 143 |
+
"eval_val_evaluator_dot_ndcg@5": 0.7265352392347777,
|
| 144 |
+
"eval_val_evaluator_dot_precision@1": 0.4846547314578005,
|
| 145 |
+
"eval_val_evaluator_dot_precision@10": 0.09616368286445011,
|
| 146 |
+
"eval_val_evaluator_dot_precision@5": 0.18439897698209717,
|
| 147 |
+
"eval_val_evaluator_dot_recall@1": 0.4846547314578005,
|
| 148 |
+
"eval_val_evaluator_dot_recall@10": 0.9616368286445013,
|
| 149 |
+
"eval_val_evaluator_dot_recall@5": 0.921994884910486,
|
| 150 |
+
"step": 1500
|
| 151 |
+
},
|
| 152 |
+
{
|
| 153 |
+
"epoch": 9.090909090909092,
|
| 154 |
+
"grad_norm": 4.445181846618652,
|
| 155 |
+
"learning_rate": 8.585858585858587e-06,
|
| 156 |
+
"loss": 0.0361,
|
| 157 |
+
"step": 2000
|
| 158 |
+
},
|
| 159 |
+
{
|
| 160 |
+
"epoch": 9.090909090909092,
|
| 161 |
+
"eval_loss": 0.06154213100671768,
|
| 162 |
+
"eval_runtime": 16.1225,
|
| 163 |
+
"eval_samples_per_second": 48.504,
|
| 164 |
+
"eval_steps_per_second": 6.078,
|
| 165 |
+
"eval_val_evaluator_cosine_accuracy@1": 0.4859335038363171,
|
| 166 |
+
"eval_val_evaluator_cosine_accuracy@10": 0.9654731457800512,
|
| 167 |
+
"eval_val_evaluator_cosine_accuracy@5": 0.9271099744245525,
|
| 168 |
+
"eval_val_evaluator_cosine_map@100": 0.6706887408572961,
|
| 169 |
+
"eval_val_evaluator_cosine_mrr@10": 0.6692648073722237,
|
| 170 |
+
"eval_val_evaluator_cosine_mrr@100": 0.6706887408572959,
|
| 171 |
+
"eval_val_evaluator_cosine_mrr@5": 0.6639812446717808,
|
| 172 |
+
"eval_val_evaluator_cosine_ndcg@10": 0.7429450115476419,
|
| 173 |
+
"eval_val_evaluator_cosine_ndcg@100": 0.749838617287201,
|
| 174 |
+
"eval_val_evaluator_cosine_ndcg@5": 0.7303784552891699,
|
| 175 |
+
"eval_val_evaluator_cosine_precision@1": 0.4859335038363171,
|
| 176 |
+
"eval_val_evaluator_cosine_precision@10": 0.09654731457800511,
|
| 177 |
+
"eval_val_evaluator_cosine_precision@5": 0.18542199488491048,
|
| 178 |
+
"eval_val_evaluator_cosine_recall@1": 0.4859335038363171,
|
| 179 |
+
"eval_val_evaluator_cosine_recall@10": 0.9654731457800512,
|
| 180 |
+
"eval_val_evaluator_cosine_recall@5": 0.9271099744245525,
|
| 181 |
+
"eval_val_evaluator_dot_accuracy@1": 0.4859335038363171,
|
| 182 |
+
"eval_val_evaluator_dot_accuracy@10": 0.9654731457800512,
|
| 183 |
+
"eval_val_evaluator_dot_accuracy@5": 0.9271099744245525,
|
| 184 |
+
"eval_val_evaluator_dot_map@100": 0.6706887408572961,
|
| 185 |
+
"eval_val_evaluator_dot_mrr@10": 0.6692648073722237,
|
| 186 |
+
"eval_val_evaluator_dot_mrr@100": 0.6706887408572959,
|
| 187 |
+
"eval_val_evaluator_dot_mrr@5": 0.6639812446717808,
|
| 188 |
+
"eval_val_evaluator_dot_ndcg@10": 0.7429450115476419,
|
| 189 |
+
"eval_val_evaluator_dot_ndcg@100": 0.749838617287201,
|
| 190 |
+
"eval_val_evaluator_dot_ndcg@5": 0.7303784552891699,
|
| 191 |
+
"eval_val_evaluator_dot_precision@1": 0.4859335038363171,
|
| 192 |
+
"eval_val_evaluator_dot_precision@10": 0.09654731457800511,
|
| 193 |
+
"eval_val_evaluator_dot_precision@5": 0.18542199488491048,
|
| 194 |
+
"eval_val_evaluator_dot_recall@1": 0.4859335038363171,
|
| 195 |
+
"eval_val_evaluator_dot_recall@10": 0.9654731457800512,
|
| 196 |
+
"eval_val_evaluator_dot_recall@5": 0.9271099744245525,
|
| 197 |
+
"step": 2000
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 11.363636363636363,
|
| 201 |
+
"grad_norm": 7.720225811004639,
|
| 202 |
+
"learning_rate": 7.954545454545455e-06,
|
| 203 |
+
"loss": 0.0301,
|
| 204 |
+
"step": 2500
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 11.363636363636363,
|
| 208 |
+
"eval_loss": 0.06867850571870804,
|
| 209 |
+
"eval_runtime": 16.1559,
|
| 210 |
+
"eval_samples_per_second": 48.403,
|
| 211 |
+
"eval_steps_per_second": 6.066,
|
| 212 |
+
"eval_val_evaluator_cosine_accuracy@1": 0.49232736572890023,
|
| 213 |
+
"eval_val_evaluator_cosine_accuracy@10": 0.9603580562659847,
|
| 214 |
+
"eval_val_evaluator_cosine_accuracy@5": 0.928388746803069,
|
| 215 |
+
"eval_val_evaluator_cosine_map@100": 0.6764517431036522,
|
| 216 |
+
"eval_val_evaluator_cosine_mrr@10": 0.6745762797872769,
|
| 217 |
+
"eval_val_evaluator_cosine_mrr@100": 0.6764517431036515,
|
| 218 |
+
"eval_val_evaluator_cosine_mrr@5": 0.6698635976129574,
|
| 219 |
+
"eval_val_evaluator_cosine_ndcg@10": 0.7459669546464727,
|
| 220 |
+
"eval_val_evaluator_cosine_ndcg@100": 0.7541025758659824,
|
| 221 |
+
"eval_val_evaluator_cosine_ndcg@5": 0.7351837372969686,
|
| 222 |
+
"eval_val_evaluator_cosine_precision@1": 0.49232736572890023,
|
| 223 |
+
"eval_val_evaluator_cosine_precision@10": 0.09603580562659846,
|
| 224 |
+
"eval_val_evaluator_cosine_precision@5": 0.1856777493606138,
|
| 225 |
+
"eval_val_evaluator_cosine_recall@1": 0.49232736572890023,
|
| 226 |
+
"eval_val_evaluator_cosine_recall@10": 0.9603580562659847,
|
| 227 |
+
"eval_val_evaluator_cosine_recall@5": 0.928388746803069,
|
| 228 |
+
"eval_val_evaluator_dot_accuracy@1": 0.49232736572890023,
|
| 229 |
+
"eval_val_evaluator_dot_accuracy@10": 0.9603580562659847,
|
| 230 |
+
"eval_val_evaluator_dot_accuracy@5": 0.928388746803069,
|
| 231 |
+
"eval_val_evaluator_dot_map@100": 0.6764517431036522,
|
| 232 |
+
"eval_val_evaluator_dot_mrr@10": 0.6745762797872769,
|
| 233 |
+
"eval_val_evaluator_dot_mrr@100": 0.6764517431036515,
|
| 234 |
+
"eval_val_evaluator_dot_mrr@5": 0.6698635976129574,
|
| 235 |
+
"eval_val_evaluator_dot_ndcg@10": 0.7459669546464727,
|
| 236 |
+
"eval_val_evaluator_dot_ndcg@100": 0.7541025758659824,
|
| 237 |
+
"eval_val_evaluator_dot_ndcg@5": 0.7351837372969686,
|
| 238 |
+
"eval_val_evaluator_dot_precision@1": 0.49232736572890023,
|
| 239 |
+
"eval_val_evaluator_dot_precision@10": 0.09603580562659846,
|
| 240 |
+
"eval_val_evaluator_dot_precision@5": 0.1856777493606138,
|
| 241 |
+
"eval_val_evaluator_dot_recall@1": 0.49232736572890023,
|
| 242 |
+
"eval_val_evaluator_dot_recall@10": 0.9603580562659847,
|
| 243 |
+
"eval_val_evaluator_dot_recall@5": 0.928388746803069,
|
| 244 |
+
"step": 2500
|
| 245 |
+
},
|
| 246 |
+
{
|
| 247 |
+
"epoch": 13.636363636363637,
|
| 248 |
+
"grad_norm": 3.5265557765960693,
|
| 249 |
+
"learning_rate": 7.323232323232324e-06,
|
| 250 |
+
"loss": 0.0274,
|
| 251 |
+
"step": 3000
|
| 252 |
+
},
|
| 253 |
+
{
|
| 254 |
+
"epoch": 13.636363636363637,
|
| 255 |
+
"eval_loss": 0.06610127538442612,
|
| 256 |
+
"eval_runtime": 16.1447,
|
| 257 |
+
"eval_samples_per_second": 48.437,
|
| 258 |
+
"eval_steps_per_second": 6.07,
|
| 259 |
+
"eval_val_evaluator_cosine_accuracy@1": 0.48721227621483376,
|
| 260 |
+
"eval_val_evaluator_cosine_accuracy@10": 0.9603580562659847,
|
| 261 |
+
"eval_val_evaluator_cosine_accuracy@5": 0.9245524296675192,
|
| 262 |
+
"eval_val_evaluator_cosine_map@100": 0.6732921649074852,
|
| 263 |
+
"eval_val_evaluator_cosine_mrr@10": 0.671490987699427,
|
| 264 |
+
"eval_val_evaluator_cosine_mrr@100": 0.6732921649074848,
|
| 265 |
+
"eval_val_evaluator_cosine_mrr@5": 0.6664748508098882,
|
| 266 |
+
"eval_val_evaluator_cosine_ndcg@10": 0.7435664719550824,
|
| 267 |
+
"eval_val_evaluator_cosine_ndcg@100": 0.7519530408382193,
|
| 268 |
+
"eval_val_evaluator_cosine_ndcg@5": 0.7317557792940828,
|
| 269 |
+
"eval_val_evaluator_cosine_precision@1": 0.48721227621483376,
|
| 270 |
+
"eval_val_evaluator_cosine_precision@10": 0.09603580562659846,
|
| 271 |
+
"eval_val_evaluator_cosine_precision@5": 0.18491048593350382,
|
| 272 |
+
"eval_val_evaluator_cosine_recall@1": 0.48721227621483376,
|
| 273 |
+
"eval_val_evaluator_cosine_recall@10": 0.9603580562659847,
|
| 274 |
+
"eval_val_evaluator_cosine_recall@5": 0.9245524296675192,
|
| 275 |
+
"eval_val_evaluator_dot_accuracy@1": 0.48721227621483376,
|
| 276 |
+
"eval_val_evaluator_dot_accuracy@10": 0.9603580562659847,
|
| 277 |
+
"eval_val_evaluator_dot_accuracy@5": 0.9245524296675192,
|
| 278 |
+
"eval_val_evaluator_dot_map@100": 0.6732921649074852,
|
| 279 |
+
"eval_val_evaluator_dot_mrr@10": 0.671490987699427,
|
| 280 |
+
"eval_val_evaluator_dot_mrr@100": 0.6732921649074848,
|
| 281 |
+
"eval_val_evaluator_dot_mrr@5": 0.6664748508098882,
|
| 282 |
+
"eval_val_evaluator_dot_ndcg@10": 0.7435664719550824,
|
| 283 |
+
"eval_val_evaluator_dot_ndcg@100": 0.7519530408382193,
|
| 284 |
+
"eval_val_evaluator_dot_ndcg@5": 0.7317557792940828,
|
| 285 |
+
"eval_val_evaluator_dot_precision@1": 0.48721227621483376,
|
| 286 |
+
"eval_val_evaluator_dot_precision@10": 0.09603580562659846,
|
| 287 |
+
"eval_val_evaluator_dot_precision@5": 0.18491048593350382,
|
| 288 |
+
"eval_val_evaluator_dot_recall@1": 0.48721227621483376,
|
| 289 |
+
"eval_val_evaluator_dot_recall@10": 0.9603580562659847,
|
| 290 |
+
"eval_val_evaluator_dot_recall@5": 0.9245524296675192,
|
| 291 |
+
"step": 3000
|
| 292 |
+
},
|
| 293 |
+
{
|
| 294 |
+
"epoch": 15.909090909090908,
|
| 295 |
+
"grad_norm": 2.537912130355835,
|
| 296 |
+
"learning_rate": 6.691919191919193e-06,
|
| 297 |
+
"loss": 0.0223,
|
| 298 |
+
"step": 3500
|
| 299 |
+
},
|
| 300 |
+
{
|
| 301 |
+
"epoch": 15.909090909090908,
|
| 302 |
+
"eval_loss": 0.06057285517454147,
|
| 303 |
+
"eval_runtime": 16.2937,
|
| 304 |
+
"eval_samples_per_second": 47.994,
|
| 305 |
+
"eval_steps_per_second": 6.015,
|
| 306 |
+
"eval_val_evaluator_cosine_accuracy@1": 0.5051150895140665,
|
| 307 |
+
"eval_val_evaluator_cosine_accuracy@10": 0.959079283887468,
|
| 308 |
+
"eval_val_evaluator_cosine_accuracy@5": 0.928388746803069,
|
| 309 |
+
"eval_val_evaluator_cosine_map@100": 0.6822101368749947,
|
| 310 |
+
"eval_val_evaluator_cosine_mrr@10": 0.680220131530873,
|
| 311 |
+
"eval_val_evaluator_cosine_mrr@100": 0.6822101368749943,
|
| 312 |
+
"eval_val_evaluator_cosine_mrr@5": 0.6759164535379364,
|
| 313 |
+
"eval_val_evaluator_cosine_ndcg@10": 0.749745056809794,
|
| 314 |
+
"eval_val_evaluator_cosine_ndcg@100": 0.7586065473558236,
|
| 315 |
+
"eval_val_evaluator_cosine_ndcg@5": 0.7396139266559898,
|
| 316 |
+
"eval_val_evaluator_cosine_precision@1": 0.5051150895140665,
|
| 317 |
+
"eval_val_evaluator_cosine_precision@10": 0.0959079283887468,
|
| 318 |
+
"eval_val_evaluator_cosine_precision@5": 0.1856777493606138,
|
| 319 |
+
"eval_val_evaluator_cosine_recall@1": 0.5051150895140665,
|
| 320 |
+
"eval_val_evaluator_cosine_recall@10": 0.959079283887468,
|
| 321 |
+
"eval_val_evaluator_cosine_recall@5": 0.928388746803069,
|
| 322 |
+
"eval_val_evaluator_dot_accuracy@1": 0.5051150895140665,
|
| 323 |
+
"eval_val_evaluator_dot_accuracy@10": 0.959079283887468,
|
| 324 |
+
"eval_val_evaluator_dot_accuracy@5": 0.928388746803069,
|
| 325 |
+
"eval_val_evaluator_dot_map@100": 0.6822101368749947,
|
| 326 |
+
"eval_val_evaluator_dot_mrr@10": 0.680220131530873,
|
| 327 |
+
"eval_val_evaluator_dot_mrr@100": 0.6822101368749943,
|
| 328 |
+
"eval_val_evaluator_dot_mrr@5": 0.6759164535379364,
|
| 329 |
+
"eval_val_evaluator_dot_ndcg@10": 0.749745056809794,
|
| 330 |
+
"eval_val_evaluator_dot_ndcg@100": 0.7586065473558236,
|
| 331 |
+
"eval_val_evaluator_dot_ndcg@5": 0.7396139266559898,
|
| 332 |
+
"eval_val_evaluator_dot_precision@1": 0.5051150895140665,
|
| 333 |
+
"eval_val_evaluator_dot_precision@10": 0.0959079283887468,
|
| 334 |
+
"eval_val_evaluator_dot_precision@5": 0.1856777493606138,
|
| 335 |
+
"eval_val_evaluator_dot_recall@1": 0.5051150895140665,
|
| 336 |
+
"eval_val_evaluator_dot_recall@10": 0.959079283887468,
|
| 337 |
+
"eval_val_evaluator_dot_recall@5": 0.928388746803069,
|
| 338 |
+
"step": 3500
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 18.181818181818183,
|
| 342 |
+
"grad_norm": 6.390940189361572,
|
| 343 |
+
"learning_rate": 6.060606060606061e-06,
|
| 344 |
+
"loss": 0.021,
|
| 345 |
+
"step": 4000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 18.181818181818183,
|
| 349 |
+
"eval_loss": 0.05628308653831482,
|
| 350 |
+
"eval_runtime": 16.1509,
|
| 351 |
+
"eval_samples_per_second": 48.418,
|
| 352 |
+
"eval_steps_per_second": 6.068,
|
| 353 |
+
"eval_val_evaluator_cosine_accuracy@1": 0.5063938618925832,
|
| 354 |
+
"eval_val_evaluator_cosine_accuracy@10": 0.9552429667519181,
|
| 355 |
+
"eval_val_evaluator_cosine_accuracy@5": 0.9245524296675192,
|
| 356 |
+
"eval_val_evaluator_cosine_map@100": 0.6833838272532422,
|
| 357 |
+
"eval_val_evaluator_cosine_mrr@10": 0.6811883449031781,
|
| 358 |
+
"eval_val_evaluator_cosine_mrr@100": 0.6833838272532414,
|
| 359 |
+
"eval_val_evaluator_cosine_mrr@5": 0.6767689684569473,
|
| 360 |
+
"eval_val_evaluator_cosine_ndcg@10": 0.7496397009089436,
|
| 361 |
+
"eval_val_evaluator_cosine_ndcg@100": 0.7591798763800799,
|
| 362 |
+
"eval_val_evaluator_cosine_ndcg@5": 0.7393934605063354,
|
| 363 |
+
"eval_val_evaluator_cosine_precision@1": 0.5063938618925832,
|
| 364 |
+
"eval_val_evaluator_cosine_precision@10": 0.0955242966751918,
|
| 365 |
+
"eval_val_evaluator_cosine_precision@5": 0.1849104859335038,
|
| 366 |
+
"eval_val_evaluator_cosine_recall@1": 0.5063938618925832,
|
| 367 |
+
"eval_val_evaluator_cosine_recall@10": 0.9552429667519181,
|
| 368 |
+
"eval_val_evaluator_cosine_recall@5": 0.9245524296675192,
|
| 369 |
+
"eval_val_evaluator_dot_accuracy@1": 0.5063938618925832,
|
| 370 |
+
"eval_val_evaluator_dot_accuracy@10": 0.9552429667519181,
|
| 371 |
+
"eval_val_evaluator_dot_accuracy@5": 0.9245524296675192,
|
| 372 |
+
"eval_val_evaluator_dot_map@100": 0.6833838272532422,
|
| 373 |
+
"eval_val_evaluator_dot_mrr@10": 0.6811883449031781,
|
| 374 |
+
"eval_val_evaluator_dot_mrr@100": 0.6833838272532414,
|
| 375 |
+
"eval_val_evaluator_dot_mrr@5": 0.6767689684569473,
|
| 376 |
+
"eval_val_evaluator_dot_ndcg@10": 0.7496397009089436,
|
| 377 |
+
"eval_val_evaluator_dot_ndcg@100": 0.7591798763800799,
|
| 378 |
+
"eval_val_evaluator_dot_ndcg@5": 0.7393934605063354,
|
| 379 |
+
"eval_val_evaluator_dot_precision@1": 0.5063938618925832,
|
| 380 |
+
"eval_val_evaluator_dot_precision@10": 0.0955242966751918,
|
| 381 |
+
"eval_val_evaluator_dot_precision@5": 0.1849104859335038,
|
| 382 |
+
"eval_val_evaluator_dot_recall@1": 0.5063938618925832,
|
| 383 |
+
"eval_val_evaluator_dot_recall@10": 0.9552429667519181,
|
| 384 |
+
"eval_val_evaluator_dot_recall@5": 0.9245524296675192,
|
| 385 |
+
"step": 4000
|
| 386 |
+
},
|
| 387 |
+
{
|
| 388 |
+
"epoch": 20.454545454545453,
|
| 389 |
+
"grad_norm": 4.392763614654541,
|
| 390 |
+
"learning_rate": 5.429292929292929e-06,
|
| 391 |
+
"loss": 0.0203,
|
| 392 |
+
"step": 4500
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 20.454545454545453,
|
| 396 |
+
"eval_loss": 0.057314660400152206,
|
| 397 |
+
"eval_runtime": 16.1911,
|
| 398 |
+
"eval_samples_per_second": 48.298,
|
| 399 |
+
"eval_steps_per_second": 6.053,
|
| 400 |
+
"eval_val_evaluator_cosine_accuracy@1": 0.4782608695652174,
|
| 401 |
+
"eval_val_evaluator_cosine_accuracy@10": 0.9578005115089514,
|
| 402 |
+
"eval_val_evaluator_cosine_accuracy@5": 0.928388746803069,
|
| 403 |
+
"eval_val_evaluator_cosine_map@100": 0.6680531566193644,
|
| 404 |
+
"eval_val_evaluator_cosine_mrr@10": 0.6661059351276735,
|
| 405 |
+
"eval_val_evaluator_cosine_mrr@100": 0.6680531566193635,
|
| 406 |
+
"eval_val_evaluator_cosine_mrr@5": 0.6618925831202036,
|
| 407 |
+
"eval_val_evaluator_cosine_ndcg@10": 0.7390313134229293,
|
| 408 |
+
"eval_val_evaluator_cosine_ndcg@100": 0.7478157002187452,
|
| 409 |
+
"eval_val_evaluator_cosine_ndcg@5": 0.7292375411745403,
|
| 410 |
+
"eval_val_evaluator_cosine_precision@1": 0.4782608695652174,
|
| 411 |
+
"eval_val_evaluator_cosine_precision@10": 0.09578005115089515,
|
| 412 |
+
"eval_val_evaluator_cosine_precision@5": 0.1856777493606138,
|
| 413 |
+
"eval_val_evaluator_cosine_recall@1": 0.4782608695652174,
|
| 414 |
+
"eval_val_evaluator_cosine_recall@10": 0.9578005115089514,
|
| 415 |
+
"eval_val_evaluator_cosine_recall@5": 0.928388746803069,
|
| 416 |
+
"eval_val_evaluator_dot_accuracy@1": 0.4782608695652174,
|
| 417 |
+
"eval_val_evaluator_dot_accuracy@10": 0.9578005115089514,
|
| 418 |
+
"eval_val_evaluator_dot_accuracy@5": 0.928388746803069,
|
| 419 |
+
"eval_val_evaluator_dot_map@100": 0.6680531566193644,
|
| 420 |
+
"eval_val_evaluator_dot_mrr@10": 0.6661059351276735,
|
| 421 |
+
"eval_val_evaluator_dot_mrr@100": 0.6680531566193635,
|
| 422 |
+
"eval_val_evaluator_dot_mrr@5": 0.6618925831202036,
|
| 423 |
+
"eval_val_evaluator_dot_ndcg@10": 0.7390313134229293,
|
| 424 |
+
"eval_val_evaluator_dot_ndcg@100": 0.7478157002187452,
|
| 425 |
+
"eval_val_evaluator_dot_ndcg@5": 0.7292375411745403,
|
| 426 |
+
"eval_val_evaluator_dot_precision@1": 0.4782608695652174,
|
| 427 |
+
"eval_val_evaluator_dot_precision@10": 0.09578005115089515,
|
| 428 |
+
"eval_val_evaluator_dot_precision@5": 0.1856777493606138,
|
| 429 |
+
"eval_val_evaluator_dot_recall@1": 0.4782608695652174,
|
| 430 |
+
"eval_val_evaluator_dot_recall@10": 0.9578005115089514,
|
| 431 |
+
"eval_val_evaluator_dot_recall@5": 0.928388746803069,
|
| 432 |
+
"step": 4500
|
| 433 |
+
},
|
| 434 |
+
{
|
| 435 |
+
"epoch": 22.727272727272727,
|
| 436 |
+
"grad_norm": 4.935870170593262,
|
| 437 |
+
"learning_rate": 4.7979797979797985e-06,
|
| 438 |
+
"loss": 0.0212,
|
| 439 |
+
"step": 5000
|
| 440 |
+
},
|
| 441 |
+
{
|
| 442 |
+
"epoch": 22.727272727272727,
|
| 443 |
+
"eval_loss": 0.06370092183351517,
|
| 444 |
+
"eval_runtime": 16.1455,
|
| 445 |
+
"eval_samples_per_second": 48.435,
|
| 446 |
+
"eval_steps_per_second": 6.07,
|
| 447 |
+
"eval_val_evaluator_cosine_accuracy@1": 0.49872122762148335,
|
| 448 |
+
"eval_val_evaluator_cosine_accuracy@10": 0.9514066496163683,
|
| 449 |
+
"eval_val_evaluator_cosine_accuracy@5": 0.9258312020460358,
|
| 450 |
+
"eval_val_evaluator_cosine_map@100": 0.6769910008307545,
|
| 451 |
+
"eval_val_evaluator_cosine_mrr@10": 0.6745772946859899,
|
| 452 |
+
"eval_val_evaluator_cosine_mrr@100": 0.6769910008307539,
|
| 453 |
+
"eval_val_evaluator_cosine_mrr@5": 0.6707587382779191,
|
| 454 |
+
"eval_val_evaluator_cosine_ndcg@10": 0.743803897516351,
|
| 455 |
+
"eval_val_evaluator_cosine_ndcg@100": 0.7540671552876991,
|
| 456 |
+
"eval_val_evaluator_cosine_ndcg@5": 0.7351253519366493,
|
| 457 |
+
"eval_val_evaluator_cosine_precision@1": 0.49872122762148335,
|
| 458 |
+
"eval_val_evaluator_cosine_precision@10": 0.0951406649616368,
|
| 459 |
+
"eval_val_evaluator_cosine_precision@5": 0.18516624040920715,
|
| 460 |
+
"eval_val_evaluator_cosine_recall@1": 0.49872122762148335,
|
| 461 |
+
"eval_val_evaluator_cosine_recall@10": 0.9514066496163683,
|
| 462 |
+
"eval_val_evaluator_cosine_recall@5": 0.9258312020460358,
|
| 463 |
+
"eval_val_evaluator_dot_accuracy@1": 0.49872122762148335,
|
| 464 |
+
"eval_val_evaluator_dot_accuracy@10": 0.9514066496163683,
|
| 465 |
+
"eval_val_evaluator_dot_accuracy@5": 0.9258312020460358,
|
| 466 |
+
"eval_val_evaluator_dot_map@100": 0.6769910008307545,
|
| 467 |
+
"eval_val_evaluator_dot_mrr@10": 0.6745772946859899,
|
| 468 |
+
"eval_val_evaluator_dot_mrr@100": 0.6769910008307539,
|
| 469 |
+
"eval_val_evaluator_dot_mrr@5": 0.6707587382779191,
|
| 470 |
+
"eval_val_evaluator_dot_ndcg@10": 0.743803897516351,
|
| 471 |
+
"eval_val_evaluator_dot_ndcg@100": 0.7540671552876991,
|
| 472 |
+
"eval_val_evaluator_dot_ndcg@5": 0.7351253519366493,
|
| 473 |
+
"eval_val_evaluator_dot_precision@1": 0.49872122762148335,
|
| 474 |
+
"eval_val_evaluator_dot_precision@10": 0.0951406649616368,
|
| 475 |
+
"eval_val_evaluator_dot_precision@5": 0.18516624040920715,
|
| 476 |
+
"eval_val_evaluator_dot_recall@1": 0.49872122762148335,
|
| 477 |
+
"eval_val_evaluator_dot_recall@10": 0.9514066496163683,
|
| 478 |
+
"eval_val_evaluator_dot_recall@5": 0.9258312020460358,
|
| 479 |
+
"step": 5000
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 25.0,
|
| 483 |
+
"grad_norm": 8.184098243713379,
|
| 484 |
+
"learning_rate": 4.166666666666667e-06,
|
| 485 |
+
"loss": 0.018,
|
| 486 |
+
"step": 5500
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 25.0,
|
| 490 |
+
"eval_loss": 0.057994794100522995,
|
| 491 |
+
"eval_runtime": 16.1282,
|
| 492 |
+
"eval_samples_per_second": 48.487,
|
| 493 |
+
"eval_steps_per_second": 6.076,
|
| 494 |
+
"eval_val_evaluator_cosine_accuracy@1": 0.49744245524296676,
|
| 495 |
+
"eval_val_evaluator_cosine_accuracy@10": 0.9539641943734015,
|
| 496 |
+
"eval_val_evaluator_cosine_accuracy@5": 0.9245524296675192,
|
| 497 |
+
"eval_val_evaluator_cosine_map@100": 0.6780504881130015,
|
| 498 |
+
"eval_val_evaluator_cosine_mrr@10": 0.6757890837494416,
|
| 499 |
+
"eval_val_evaluator_cosine_mrr@100": 0.6780504881130013,
|
| 500 |
+
"eval_val_evaluator_cosine_mrr@5": 0.6714407502131281,
|
| 501 |
+
"eval_val_evaluator_cosine_ndcg@10": 0.7453518958660976,
|
| 502 |
+
"eval_val_evaluator_cosine_ndcg@100": 0.7550072905243007,
|
| 503 |
+
"eval_val_evaluator_cosine_ndcg@5": 0.7354200452387065,
|
| 504 |
+
"eval_val_evaluator_cosine_precision@1": 0.49744245524296676,
|
| 505 |
+
"eval_val_evaluator_cosine_precision@10": 0.09539641943734015,
|
| 506 |
+
"eval_val_evaluator_cosine_precision@5": 0.18491048593350382,
|
| 507 |
+
"eval_val_evaluator_cosine_recall@1": 0.49744245524296676,
|
| 508 |
+
"eval_val_evaluator_cosine_recall@10": 0.9539641943734015,
|
| 509 |
+
"eval_val_evaluator_cosine_recall@5": 0.9245524296675192,
|
| 510 |
+
"eval_val_evaluator_dot_accuracy@1": 0.49744245524296676,
|
| 511 |
+
"eval_val_evaluator_dot_accuracy@10": 0.9539641943734015,
|
| 512 |
+
"eval_val_evaluator_dot_accuracy@5": 0.9245524296675192,
|
| 513 |
+
"eval_val_evaluator_dot_map@100": 0.6780504881130015,
|
| 514 |
+
"eval_val_evaluator_dot_mrr@10": 0.6757890837494416,
|
| 515 |
+
"eval_val_evaluator_dot_mrr@100": 0.6780504881130013,
|
| 516 |
+
"eval_val_evaluator_dot_mrr@5": 0.6714407502131281,
|
| 517 |
+
"eval_val_evaluator_dot_ndcg@10": 0.7453518958660976,
|
| 518 |
+
"eval_val_evaluator_dot_ndcg@100": 0.7550072905243007,
|
| 519 |
+
"eval_val_evaluator_dot_ndcg@5": 0.7354200452387065,
|
| 520 |
+
"eval_val_evaluator_dot_precision@1": 0.49744245524296676,
|
| 521 |
+
"eval_val_evaluator_dot_precision@10": 0.09539641943734015,
|
| 522 |
+
"eval_val_evaluator_dot_precision@5": 0.18491048593350382,
|
| 523 |
+
"eval_val_evaluator_dot_recall@1": 0.49744245524296676,
|
| 524 |
+
"eval_val_evaluator_dot_recall@10": 0.9539641943734015,
|
| 525 |
+
"eval_val_evaluator_dot_recall@5": 0.9245524296675192,
|
| 526 |
+
"step": 5500
|
| 527 |
+
},
|
| 528 |
+
{
|
| 529 |
+
"epoch": 27.272727272727273,
|
| 530 |
+
"grad_norm": 0.004099330864846706,
|
| 531 |
+
"learning_rate": 3.5353535353535356e-06,
|
| 532 |
+
"loss": 0.0166,
|
| 533 |
+
"step": 6000
|
| 534 |
+
},
|
| 535 |
+
{
|
| 536 |
+
"epoch": 27.272727272727273,
|
| 537 |
+
"eval_loss": 0.056713227182626724,
|
| 538 |
+
"eval_runtime": 16.1679,
|
| 539 |
+
"eval_samples_per_second": 48.367,
|
| 540 |
+
"eval_steps_per_second": 6.061,
|
| 541 |
+
"eval_val_evaluator_cosine_accuracy@1": 0.4948849104859335,
|
| 542 |
+
"eval_val_evaluator_cosine_accuracy@10": 0.9552429667519181,
|
| 543 |
+
"eval_val_evaluator_cosine_accuracy@5": 0.9232736572890026,
|
| 544 |
+
"eval_val_evaluator_cosine_map@100": 0.6780550124960388,
|
| 545 |
+
"eval_val_evaluator_cosine_mrr@10": 0.675847440425445,
|
| 546 |
+
"eval_val_evaluator_cosine_mrr@100": 0.6780550124960384,
|
| 547 |
+
"eval_val_evaluator_cosine_mrr@5": 0.6710358056265976,
|
| 548 |
+
"eval_val_evaluator_cosine_ndcg@10": 0.7457753760186268,
|
| 549 |
+
"eval_val_evaluator_cosine_ndcg@100": 0.7551473066673557,
|
| 550 |
+
"eval_val_evaluator_cosine_ndcg@5": 0.7348943591506862,
|
| 551 |
+
"eval_val_evaluator_cosine_precision@1": 0.4948849104859335,
|
| 552 |
+
"eval_val_evaluator_cosine_precision@10": 0.0955242966751918,
|
| 553 |
+
"eval_val_evaluator_cosine_precision@5": 0.18465473145780048,
|
| 554 |
+
"eval_val_evaluator_cosine_recall@1": 0.4948849104859335,
|
| 555 |
+
"eval_val_evaluator_cosine_recall@10": 0.9552429667519181,
|
| 556 |
+
"eval_val_evaluator_cosine_recall@5": 0.9232736572890026,
|
| 557 |
+
"eval_val_evaluator_dot_accuracy@1": 0.4948849104859335,
|
| 558 |
+
"eval_val_evaluator_dot_accuracy@10": 0.9552429667519181,
|
| 559 |
+
"eval_val_evaluator_dot_accuracy@5": 0.9232736572890026,
|
| 560 |
+
"eval_val_evaluator_dot_map@100": 0.6780550124960388,
|
| 561 |
+
"eval_val_evaluator_dot_mrr@10": 0.675847440425445,
|
| 562 |
+
"eval_val_evaluator_dot_mrr@100": 0.6780550124960384,
|
| 563 |
+
"eval_val_evaluator_dot_mrr@5": 0.6710358056265976,
|
| 564 |
+
"eval_val_evaluator_dot_ndcg@10": 0.7457753760186268,
|
| 565 |
+
"eval_val_evaluator_dot_ndcg@100": 0.7551473066673557,
|
| 566 |
+
"eval_val_evaluator_dot_ndcg@5": 0.7348943591506862,
|
| 567 |
+
"eval_val_evaluator_dot_precision@1": 0.4948849104859335,
|
| 568 |
+
"eval_val_evaluator_dot_precision@10": 0.0955242966751918,
|
| 569 |
+
"eval_val_evaluator_dot_precision@5": 0.18465473145780048,
|
| 570 |
+
"eval_val_evaluator_dot_recall@1": 0.4948849104859335,
|
| 571 |
+
"eval_val_evaluator_dot_recall@10": 0.9552429667519181,
|
| 572 |
+
"eval_val_evaluator_dot_recall@5": 0.9232736572890026,
|
| 573 |
+
"step": 6000
|
| 574 |
+
},
|
| 575 |
+
{
|
| 576 |
+
"epoch": 29.545454545454547,
|
| 577 |
+
"grad_norm": 0.02236221544444561,
|
| 578 |
+
"learning_rate": 2.904040404040404e-06,
|
| 579 |
+
"loss": 0.0194,
|
| 580 |
+
"step": 6500
|
| 581 |
+
},
|
| 582 |
+
{
|
| 583 |
+
"epoch": 29.545454545454547,
|
| 584 |
+
"eval_loss": 0.05420228838920593,
|
| 585 |
+
"eval_runtime": 16.1682,
|
| 586 |
+
"eval_samples_per_second": 48.366,
|
| 587 |
+
"eval_steps_per_second": 6.061,
|
| 588 |
+
"eval_val_evaluator_cosine_accuracy@1": 0.5038363171355499,
|
| 589 |
+
"eval_val_evaluator_cosine_accuracy@10": 0.9565217391304348,
|
| 590 |
+
"eval_val_evaluator_cosine_accuracy@5": 0.9271099744245525,
|
| 591 |
+
"eval_val_evaluator_cosine_map@100": 0.6834975753216492,
|
| 592 |
+
"eval_val_evaluator_cosine_mrr@10": 0.6813415946088577,
|
| 593 |
+
"eval_val_evaluator_cosine_mrr@100": 0.6834975753216488,
|
| 594 |
+
"eval_val_evaluator_cosine_mrr@5": 0.6769181585677743,
|
| 595 |
+
"eval_val_evaluator_cosine_ndcg@10": 0.7501703504734094,
|
| 596 |
+
"eval_val_evaluator_cosine_ndcg@100": 0.759281214123215,
|
| 597 |
+
"eval_val_evaluator_cosine_ndcg@5": 0.7401646982131597,
|
| 598 |
+
"eval_val_evaluator_cosine_precision@1": 0.5038363171355499,
|
| 599 |
+
"eval_val_evaluator_cosine_precision@10": 0.09565217391304348,
|
| 600 |
+
"eval_val_evaluator_cosine_precision@5": 0.18542199488491046,
|
| 601 |
+
"eval_val_evaluator_cosine_recall@1": 0.5038363171355499,
|
| 602 |
+
"eval_val_evaluator_cosine_recall@10": 0.9565217391304348,
|
| 603 |
+
"eval_val_evaluator_cosine_recall@5": 0.9271099744245525,
|
| 604 |
+
"eval_val_evaluator_dot_accuracy@1": 0.5038363171355499,
|
| 605 |
+
"eval_val_evaluator_dot_accuracy@10": 0.9565217391304348,
|
| 606 |
+
"eval_val_evaluator_dot_accuracy@5": 0.9271099744245525,
|
| 607 |
+
"eval_val_evaluator_dot_map@100": 0.6834975753216492,
|
| 608 |
+
"eval_val_evaluator_dot_mrr@10": 0.6813415946088577,
|
| 609 |
+
"eval_val_evaluator_dot_mrr@100": 0.6834975753216488,
|
| 610 |
+
"eval_val_evaluator_dot_mrr@5": 0.6769181585677743,
|
| 611 |
+
"eval_val_evaluator_dot_ndcg@10": 0.7501703504734094,
|
| 612 |
+
"eval_val_evaluator_dot_ndcg@100": 0.759281214123215,
|
| 613 |
+
"eval_val_evaluator_dot_ndcg@5": 0.7401646982131597,
|
| 614 |
+
"eval_val_evaluator_dot_precision@1": 0.5038363171355499,
|
| 615 |
+
"eval_val_evaluator_dot_precision@10": 0.09565217391304348,
|
| 616 |
+
"eval_val_evaluator_dot_precision@5": 0.18542199488491046,
|
| 617 |
+
"eval_val_evaluator_dot_recall@1": 0.5038363171355499,
|
| 618 |
+
"eval_val_evaluator_dot_recall@10": 0.9565217391304348,
|
| 619 |
+
"eval_val_evaluator_dot_recall@5": 0.9271099744245525,
|
| 620 |
+
"step": 6500
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 31.818181818181817,
|
| 624 |
+
"grad_norm": 0.03244137763977051,
|
| 625 |
+
"learning_rate": 2.2727272727272728e-06,
|
| 626 |
+
"loss": 0.0182,
|
| 627 |
+
"step": 7000
|
| 628 |
+
},
|
| 629 |
+
{
|
| 630 |
+
"epoch": 31.818181818181817,
|
| 631 |
+
"eval_loss": 0.05468086898326874,
|
| 632 |
+
"eval_runtime": 23.5877,
|
| 633 |
+
"eval_samples_per_second": 33.153,
|
| 634 |
+
"eval_steps_per_second": 4.155,
|
| 635 |
+
"eval_val_evaluator_cosine_accuracy@1": 0.5140664961636828,
|
| 636 |
+
"eval_val_evaluator_cosine_accuracy@10": 0.9578005115089514,
|
| 637 |
+
"eval_val_evaluator_cosine_accuracy@5": 0.9296675191815856,
|
| 638 |
+
"eval_val_evaluator_cosine_map@100": 0.6896503713996865,
|
| 639 |
+
"eval_val_evaluator_cosine_mrr@10": 0.6875806844476918,
|
| 640 |
+
"eval_val_evaluator_cosine_mrr@100": 0.6896503713996863,
|
| 641 |
+
"eval_val_evaluator_cosine_mrr@5": 0.6833333333333328,
|
| 642 |
+
"eval_val_evaluator_cosine_ndcg@10": 0.7551953164170838,
|
| 643 |
+
"eval_val_evaluator_cosine_ndcg@100": 0.7640059344945531,
|
| 644 |
+
"eval_val_evaluator_cosine_ndcg@5": 0.7456070216519929,
|
| 645 |
+
"eval_val_evaluator_cosine_precision@1": 0.5140664961636828,
|
| 646 |
+
"eval_val_evaluator_cosine_precision@10": 0.09578005115089515,
|
| 647 |
+
"eval_val_evaluator_cosine_precision@5": 0.1859335038363171,
|
| 648 |
+
"eval_val_evaluator_cosine_recall@1": 0.5140664961636828,
|
| 649 |
+
"eval_val_evaluator_cosine_recall@10": 0.9578005115089514,
|
| 650 |
+
"eval_val_evaluator_cosine_recall@5": 0.9296675191815856,
|
| 651 |
+
"eval_val_evaluator_dot_accuracy@1": 0.5140664961636828,
|
| 652 |
+
"eval_val_evaluator_dot_accuracy@10": 0.9578005115089514,
|
| 653 |
+
"eval_val_evaluator_dot_accuracy@5": 0.9296675191815856,
|
| 654 |
+
"eval_val_evaluator_dot_map@100": 0.6896503713996865,
|
| 655 |
+
"eval_val_evaluator_dot_mrr@10": 0.6875806844476918,
|
| 656 |
+
"eval_val_evaluator_dot_mrr@100": 0.6896503713996863,
|
| 657 |
+
"eval_val_evaluator_dot_mrr@5": 0.6833333333333328,
|
| 658 |
+
"eval_val_evaluator_dot_ndcg@10": 0.7551953164170838,
|
| 659 |
+
"eval_val_evaluator_dot_ndcg@100": 0.7640059344945531,
|
| 660 |
+
"eval_val_evaluator_dot_ndcg@5": 0.7456070216519929,
|
| 661 |
+
"eval_val_evaluator_dot_precision@1": 0.5140664961636828,
|
| 662 |
+
"eval_val_evaluator_dot_precision@10": 0.09578005115089515,
|
| 663 |
+
"eval_val_evaluator_dot_precision@5": 0.1859335038363171,
|
| 664 |
+
"eval_val_evaluator_dot_recall@1": 0.5140664961636828,
|
| 665 |
+
"eval_val_evaluator_dot_recall@10": 0.9578005115089514,
|
| 666 |
+
"eval_val_evaluator_dot_recall@5": 0.9296675191815856,
|
| 667 |
+
"step": 7000
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 34.09090909090909,
|
| 671 |
+
"grad_norm": 0.001623653806746006,
|
| 672 |
+
"learning_rate": 1.6414141414141415e-06,
|
| 673 |
+
"loss": 0.0157,
|
| 674 |
+
"step": 7500
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 34.09090909090909,
|
| 678 |
+
"eval_loss": 0.05486880987882614,
|
| 679 |
+
"eval_runtime": 16.1005,
|
| 680 |
+
"eval_samples_per_second": 48.57,
|
| 681 |
+
"eval_steps_per_second": 6.087,
|
| 682 |
+
"eval_val_evaluator_cosine_accuracy@1": 0.5140664961636828,
|
| 683 |
+
"eval_val_evaluator_cosine_accuracy@10": 0.9565217391304348,
|
| 684 |
+
"eval_val_evaluator_cosine_accuracy@5": 0.9309462915601023,
|
| 685 |
+
"eval_val_evaluator_cosine_map@100": 0.6898639676357106,
|
| 686 |
+
"eval_val_evaluator_cosine_mrr@10": 0.6877618438679816,
|
| 687 |
+
"eval_val_evaluator_cosine_mrr@100": 0.6898639676357101,
|
| 688 |
+
"eval_val_evaluator_cosine_mrr@5": 0.6839727195225911,
|
| 689 |
+
"eval_val_evaluator_cosine_ndcg@10": 0.7550036263650965,
|
| 690 |
+
"eval_val_evaluator_cosine_ndcg@100": 0.764061133822081,
|
| 691 |
+
"eval_val_evaluator_cosine_ndcg@5": 0.7463601070791411,
|
| 692 |
+
"eval_val_evaluator_cosine_precision@1": 0.5140664961636828,
|
| 693 |
+
"eval_val_evaluator_cosine_precision@10": 0.09565217391304345,
|
| 694 |
+
"eval_val_evaluator_cosine_precision@5": 0.18618925831202046,
|
| 695 |
+
"eval_val_evaluator_cosine_recall@1": 0.5140664961636828,
|
| 696 |
+
"eval_val_evaluator_cosine_recall@10": 0.9565217391304348,
|
| 697 |
+
"eval_val_evaluator_cosine_recall@5": 0.9309462915601023,
|
| 698 |
+
"eval_val_evaluator_dot_accuracy@1": 0.5140664961636828,
|
| 699 |
+
"eval_val_evaluator_dot_accuracy@10": 0.9565217391304348,
|
| 700 |
+
"eval_val_evaluator_dot_accuracy@5": 0.9309462915601023,
|
| 701 |
+
"eval_val_evaluator_dot_map@100": 0.6898639676357106,
|
| 702 |
+
"eval_val_evaluator_dot_mrr@10": 0.6877618438679816,
|
| 703 |
+
"eval_val_evaluator_dot_mrr@100": 0.6898639676357101,
|
| 704 |
+
"eval_val_evaluator_dot_mrr@5": 0.6839727195225911,
|
| 705 |
+
"eval_val_evaluator_dot_ndcg@10": 0.7550036263650965,
|
| 706 |
+
"eval_val_evaluator_dot_ndcg@100": 0.764061133822081,
|
| 707 |
+
"eval_val_evaluator_dot_ndcg@5": 0.7463601070791411,
|
| 708 |
+
"eval_val_evaluator_dot_precision@1": 0.5140664961636828,
|
| 709 |
+
"eval_val_evaluator_dot_precision@10": 0.09565217391304345,
|
| 710 |
+
"eval_val_evaluator_dot_precision@5": 0.18618925831202046,
|
| 711 |
+
"eval_val_evaluator_dot_recall@1": 0.5140664961636828,
|
| 712 |
+
"eval_val_evaluator_dot_recall@10": 0.9565217391304348,
|
| 713 |
+
"eval_val_evaluator_dot_recall@5": 0.9309462915601023,
|
| 714 |
+
"step": 7500
|
| 715 |
+
},
|
| 716 |
+
{
|
| 717 |
+
"epoch": 36.36363636363637,
|
| 718 |
+
"grad_norm": 0.06726043671369553,
|
| 719 |
+
"learning_rate": 1.01010101010101e-06,
|
| 720 |
+
"loss": 0.016,
|
| 721 |
+
"step": 8000
|
| 722 |
+
},
|
| 723 |
+
{
|
| 724 |
+
"epoch": 36.36363636363637,
|
| 725 |
+
"eval_loss": 0.05301735922694206,
|
| 726 |
+
"eval_runtime": 16.2548,
|
| 727 |
+
"eval_samples_per_second": 48.109,
|
| 728 |
+
"eval_steps_per_second": 6.029,
|
| 729 |
+
"eval_val_evaluator_cosine_accuracy@1": 0.5089514066496164,
|
| 730 |
+
"eval_val_evaluator_cosine_accuracy@10": 0.9578005115089514,
|
| 731 |
+
"eval_val_evaluator_cosine_accuracy@5": 0.9347826086956522,
|
| 732 |
+
"eval_val_evaluator_cosine_map@100": 0.6859605033261141,
|
| 733 |
+
"eval_val_evaluator_cosine_mrr@10": 0.6839808387122961,
|
| 734 |
+
"eval_val_evaluator_cosine_mrr@100": 0.6859605033261138,
|
| 735 |
+
"eval_val_evaluator_cosine_mrr@5": 0.6806052855924973,
|
| 736 |
+
"eval_val_evaluator_cosine_ndcg@10": 0.752452574251246,
|
| 737 |
+
"eval_val_evaluator_cosine_ndcg@100": 0.7611500008339412,
|
| 738 |
+
"eval_val_evaluator_cosine_ndcg@5": 0.7447092792639648,
|
| 739 |
+
"eval_val_evaluator_cosine_precision@1": 0.5089514066496164,
|
| 740 |
+
"eval_val_evaluator_cosine_precision@10": 0.09578005115089515,
|
| 741 |
+
"eval_val_evaluator_cosine_precision@5": 0.1869565217391304,
|
| 742 |
+
"eval_val_evaluator_cosine_recall@1": 0.5089514066496164,
|
| 743 |
+
"eval_val_evaluator_cosine_recall@10": 0.9578005115089514,
|
| 744 |
+
"eval_val_evaluator_cosine_recall@5": 0.9347826086956522,
|
| 745 |
+
"eval_val_evaluator_dot_accuracy@1": 0.5089514066496164,
|
| 746 |
+
"eval_val_evaluator_dot_accuracy@10": 0.9578005115089514,
|
| 747 |
+
"eval_val_evaluator_dot_accuracy@5": 0.9347826086956522,
|
| 748 |
+
"eval_val_evaluator_dot_map@100": 0.6859605033261141,
|
| 749 |
+
"eval_val_evaluator_dot_mrr@10": 0.6839808387122961,
|
| 750 |
+
"eval_val_evaluator_dot_mrr@100": 0.6859605033261138,
|
| 751 |
+
"eval_val_evaluator_dot_mrr@5": 0.6806052855924973,
|
| 752 |
+
"eval_val_evaluator_dot_ndcg@10": 0.752452574251246,
|
| 753 |
+
"eval_val_evaluator_dot_ndcg@100": 0.7611500008339412,
|
| 754 |
+
"eval_val_evaluator_dot_ndcg@5": 0.7447092792639648,
|
| 755 |
+
"eval_val_evaluator_dot_precision@1": 0.5089514066496164,
|
| 756 |
+
"eval_val_evaluator_dot_precision@10": 0.09578005115089515,
|
| 757 |
+
"eval_val_evaluator_dot_precision@5": 0.1869565217391304,
|
| 758 |
+
"eval_val_evaluator_dot_recall@1": 0.5089514066496164,
|
| 759 |
+
"eval_val_evaluator_dot_recall@10": 0.9578005115089514,
|
| 760 |
+
"eval_val_evaluator_dot_recall@5": 0.9347826086956522,
|
| 761 |
+
"step": 8000
|
| 762 |
+
}
|
| 763 |
+
],
|
| 764 |
+
"logging_steps": 500,
|
| 765 |
+
"max_steps": 8800,
|
| 766 |
+
"num_input_tokens_seen": 0,
|
| 767 |
+
"num_train_epochs": 40,
|
| 768 |
+
"save_steps": 500,
|
| 769 |
+
"stateful_callbacks": {
|
| 770 |
+
"TrainerControl": {
|
| 771 |
+
"args": {
|
| 772 |
+
"should_epoch_stop": false,
|
| 773 |
+
"should_evaluate": false,
|
| 774 |
+
"should_log": false,
|
| 775 |
+
"should_save": true,
|
| 776 |
+
"should_training_stop": false
|
| 777 |
+
},
|
| 778 |
+
"attributes": {}
|
| 779 |
+
}
|
| 780 |
+
},
|
| 781 |
+
"total_flos": 0.0,
|
| 782 |
+
"train_batch_size": 8,
|
| 783 |
+
"trial_name": null,
|
| 784 |
+
"trial_params": null
|
| 785 |
+
}
|
checkpoint-8000/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:47648917f49529e1cd5c4f55390eb54738fe83ae8b04b361ce299a8b142ad8a8
|
| 3 |
+
size 5304
|
checkpoint-8000/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
checkpoint-8500/1_Pooling/config.json
ADDED
|
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"word_embedding_dimension": 384,
|
| 3 |
+
"pooling_mode_cls_token": true,
|
| 4 |
+
"pooling_mode_mean_tokens": false,
|
| 5 |
+
"pooling_mode_max_tokens": false,
|
| 6 |
+
"pooling_mode_mean_sqrt_len_tokens": false,
|
| 7 |
+
"pooling_mode_weightedmean_tokens": false,
|
| 8 |
+
"pooling_mode_lasttoken": false,
|
| 9 |
+
"include_prompt": true
|
| 10 |
+
}
|
checkpoint-8500/README.md
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
checkpoint-8500/config.json
ADDED
|
@@ -0,0 +1,31 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "BAAI/bge-small-en-v1.5",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"BertModel"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.1,
|
| 7 |
+
"classifier_dropout": null,
|
| 8 |
+
"hidden_act": "gelu",
|
| 9 |
+
"hidden_dropout_prob": 0.1,
|
| 10 |
+
"hidden_size": 384,
|
| 11 |
+
"id2label": {
|
| 12 |
+
"0": "LABEL_0"
|
| 13 |
+
},
|
| 14 |
+
"initializer_range": 0.02,
|
| 15 |
+
"intermediate_size": 1536,
|
| 16 |
+
"label2id": {
|
| 17 |
+
"LABEL_0": 0
|
| 18 |
+
},
|
| 19 |
+
"layer_norm_eps": 1e-12,
|
| 20 |
+
"max_position_embeddings": 512,
|
| 21 |
+
"model_type": "bert",
|
| 22 |
+
"num_attention_heads": 12,
|
| 23 |
+
"num_hidden_layers": 12,
|
| 24 |
+
"pad_token_id": 0,
|
| 25 |
+
"position_embedding_type": "absolute",
|
| 26 |
+
"torch_dtype": "float32",
|
| 27 |
+
"transformers_version": "4.41.1",
|
| 28 |
+
"type_vocab_size": 2,
|
| 29 |
+
"use_cache": true,
|
| 30 |
+
"vocab_size": 30522
|
| 31 |
+
}
|
checkpoint-8500/config_sentence_transformers.json
ADDED
|
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"__version__": {
|
| 3 |
+
"sentence_transformers": "3.0.1",
|
| 4 |
+
"transformers": "4.41.1",
|
| 5 |
+
"pytorch": "2.3.1+cu121"
|
| 6 |
+
},
|
| 7 |
+
"prompts": {},
|
| 8 |
+
"default_prompt_name": null,
|
| 9 |
+
"similarity_fn_name": null
|
| 10 |
+
}
|
checkpoint-8500/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a1187527dff59b28f494768c851c6b76f4e1c5bb189c14f5cb509c35886e9d5a
|
| 3 |
+
size 133462128
|
checkpoint-8500/modules.json
ADDED
|
@@ -0,0 +1,20 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
[
|
| 2 |
+
{
|
| 3 |
+
"idx": 0,
|
| 4 |
+
"name": "0",
|
| 5 |
+
"path": "",
|
| 6 |
+
"type": "sentence_transformers.models.Transformer"
|
| 7 |
+
},
|
| 8 |
+
{
|
| 9 |
+
"idx": 1,
|
| 10 |
+
"name": "1",
|
| 11 |
+
"path": "1_Pooling",
|
| 12 |
+
"type": "sentence_transformers.models.Pooling"
|
| 13 |
+
},
|
| 14 |
+
{
|
| 15 |
+
"idx": 2,
|
| 16 |
+
"name": "2",
|
| 17 |
+
"path": "2_Normalize",
|
| 18 |
+
"type": "sentence_transformers.models.Normalize"
|
| 19 |
+
}
|
| 20 |
+
]
|
checkpoint-8500/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:004be7643a9c3e7e3e248169d9acd0f439129615fec1251fdbdeb9cbfb3a6270
|
| 3 |
+
size 265862074
|
checkpoint-8500/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:62cd051357630435144f7b8f08d52fb3696dec09e993e124d1931a7111cf9d69
|
| 3 |
+
size 14244
|
checkpoint-8500/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c3289b91ffcc34ca836ed25f8e0c65fe3a212eeda1f9ba8c13cbb55742e285b9
|
| 3 |
+
size 1064
|
checkpoint-8500/sentence_bert_config.json
ADDED
|
@@ -0,0 +1,4 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"max_seq_length": 512,
|
| 3 |
+
"do_lower_case": true
|
| 4 |
+
}
|
checkpoint-8500/special_tokens_map.json
ADDED
|
@@ -0,0 +1,37 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"cls_token": {
|
| 3 |
+
"content": "[CLS]",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": false,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"mask_token": {
|
| 10 |
+
"content": "[MASK]",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"pad_token": {
|
| 17 |
+
"content": "[PAD]",
|
| 18 |
+
"lstrip": false,
|
| 19 |
+
"normalized": false,
|
| 20 |
+
"rstrip": false,
|
| 21 |
+
"single_word": false
|
| 22 |
+
},
|
| 23 |
+
"sep_token": {
|
| 24 |
+
"content": "[SEP]",
|
| 25 |
+
"lstrip": false,
|
| 26 |
+
"normalized": false,
|
| 27 |
+
"rstrip": false,
|
| 28 |
+
"single_word": false
|
| 29 |
+
},
|
| 30 |
+
"unk_token": {
|
| 31 |
+
"content": "[UNK]",
|
| 32 |
+
"lstrip": false,
|
| 33 |
+
"normalized": false,
|
| 34 |
+
"rstrip": false,
|
| 35 |
+
"single_word": false
|
| 36 |
+
}
|
| 37 |
+
}
|
checkpoint-8500/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
checkpoint-8500/tokenizer_config.json
ADDED
|
@@ -0,0 +1,57 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "[PAD]",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"100": {
|
| 12 |
+
"content": "[UNK]",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"101": {
|
| 20 |
+
"content": "[CLS]",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"102": {
|
| 28 |
+
"content": "[SEP]",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"103": {
|
| 36 |
+
"content": "[MASK]",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"clean_up_tokenization_spaces": true,
|
| 45 |
+
"cls_token": "[CLS]",
|
| 46 |
+
"do_basic_tokenize": true,
|
| 47 |
+
"do_lower_case": true,
|
| 48 |
+
"mask_token": "[MASK]",
|
| 49 |
+
"model_max_length": 512,
|
| 50 |
+
"never_split": null,
|
| 51 |
+
"pad_token": "[PAD]",
|
| 52 |
+
"sep_token": "[SEP]",
|
| 53 |
+
"strip_accents": null,
|
| 54 |
+
"tokenize_chinese_chars": true,
|
| 55 |
+
"tokenizer_class": "BertTokenizer",
|
| 56 |
+
"unk_token": "[UNK]"
|
| 57 |
+
}
|
checkpoint-8500/trainer_state.json
ADDED
|
@@ -0,0 +1,832 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.05301735922694206,
|
| 3 |
+
"best_model_checkpoint": "bge_finetune_dsv2/checkpoint-8000",
|
| 4 |
+
"epoch": 38.63636363636363,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 8500,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 2.2727272727272725,
|
| 13 |
+
"grad_norm": 2.923750162124634,
|
| 14 |
+
"learning_rate": 5.681818181818183e-06,
|
| 15 |
+
"loss": 0.2767,
|
| 16 |
+
"step": 500
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 2.2727272727272725,
|
| 20 |
+
"eval_loss": 0.09305377304553986,
|
| 21 |
+
"eval_runtime": 16.1531,
|
| 22 |
+
"eval_samples_per_second": 48.412,
|
| 23 |
+
"eval_steps_per_second": 6.067,
|
| 24 |
+
"eval_val_evaluator_cosine_accuracy@1": 0.46675191815856776,
|
| 25 |
+
"eval_val_evaluator_cosine_accuracy@10": 0.9411764705882353,
|
| 26 |
+
"eval_val_evaluator_cosine_accuracy@5": 0.8913043478260869,
|
| 27 |
+
"eval_val_evaluator_cosine_map@100": 0.6449144076353992,
|
| 28 |
+
"eval_val_evaluator_cosine_mrr@10": 0.6419812852677294,
|
| 29 |
+
"eval_val_evaluator_cosine_mrr@100": 0.6449144076353982,
|
| 30 |
+
"eval_val_evaluator_cosine_mrr@5": 0.6351449275362308,
|
| 31 |
+
"eval_val_evaluator_cosine_ndcg@10": 0.715922596095415,
|
| 32 |
+
"eval_val_evaluator_cosine_ndcg@100": 0.7286582483109194,
|
| 33 |
+
"eval_val_evaluator_cosine_ndcg@5": 0.6996180454943595,
|
| 34 |
+
"eval_val_evaluator_cosine_precision@1": 0.46675191815856776,
|
| 35 |
+
"eval_val_evaluator_cosine_precision@10": 0.09411764705882353,
|
| 36 |
+
"eval_val_evaluator_cosine_precision@5": 0.17826086956521736,
|
| 37 |
+
"eval_val_evaluator_cosine_recall@1": 0.46675191815856776,
|
| 38 |
+
"eval_val_evaluator_cosine_recall@10": 0.9411764705882353,
|
| 39 |
+
"eval_val_evaluator_cosine_recall@5": 0.8913043478260869,
|
| 40 |
+
"eval_val_evaluator_dot_accuracy@1": 0.46675191815856776,
|
| 41 |
+
"eval_val_evaluator_dot_accuracy@10": 0.9411764705882353,
|
| 42 |
+
"eval_val_evaluator_dot_accuracy@5": 0.8913043478260869,
|
| 43 |
+
"eval_val_evaluator_dot_map@100": 0.6449144076353992,
|
| 44 |
+
"eval_val_evaluator_dot_mrr@10": 0.6419812852677294,
|
| 45 |
+
"eval_val_evaluator_dot_mrr@100": 0.6449144076353982,
|
| 46 |
+
"eval_val_evaluator_dot_mrr@5": 0.6351449275362308,
|
| 47 |
+
"eval_val_evaluator_dot_ndcg@10": 0.715922596095415,
|
| 48 |
+
"eval_val_evaluator_dot_ndcg@100": 0.7286582483109194,
|
| 49 |
+
"eval_val_evaluator_dot_ndcg@5": 0.6996180454943595,
|
| 50 |
+
"eval_val_evaluator_dot_precision@1": 0.46675191815856776,
|
| 51 |
+
"eval_val_evaluator_dot_precision@10": 0.09411764705882353,
|
| 52 |
+
"eval_val_evaluator_dot_precision@5": 0.17826086956521736,
|
| 53 |
+
"eval_val_evaluator_dot_recall@1": 0.46675191815856776,
|
| 54 |
+
"eval_val_evaluator_dot_recall@10": 0.9411764705882353,
|
| 55 |
+
"eval_val_evaluator_dot_recall@5": 0.8913043478260869,
|
| 56 |
+
"step": 500
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 4.545454545454545,
|
| 60 |
+
"grad_norm": 1.0960533618927002,
|
| 61 |
+
"learning_rate": 9.84848484848485e-06,
|
| 62 |
+
"loss": 0.067,
|
| 63 |
+
"step": 1000
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 4.545454545454545,
|
| 67 |
+
"eval_loss": 0.07767628133296967,
|
| 68 |
+
"eval_runtime": 16.162,
|
| 69 |
+
"eval_samples_per_second": 48.385,
|
| 70 |
+
"eval_steps_per_second": 6.064,
|
| 71 |
+
"eval_val_evaluator_cosine_accuracy@1": 0.4680306905370844,
|
| 72 |
+
"eval_val_evaluator_cosine_accuracy@10": 0.948849104859335,
|
| 73 |
+
"eval_val_evaluator_cosine_accuracy@5": 0.8951406649616368,
|
| 74 |
+
"eval_val_evaluator_cosine_map@100": 0.6501350929872222,
|
| 75 |
+
"eval_val_evaluator_cosine_mrr@10": 0.6475830187147313,
|
| 76 |
+
"eval_val_evaluator_cosine_mrr@100": 0.6501350929872212,
|
| 77 |
+
"eval_val_evaluator_cosine_mrr@5": 0.6400468883205445,
|
| 78 |
+
"eval_val_evaluator_cosine_ndcg@10": 0.7221983441504519,
|
| 79 |
+
"eval_val_evaluator_cosine_ndcg@100": 0.7332422183636672,
|
| 80 |
+
"eval_val_evaluator_cosine_ndcg@5": 0.7044596320615103,
|
| 81 |
+
"eval_val_evaluator_cosine_precision@1": 0.4680306905370844,
|
| 82 |
+
"eval_val_evaluator_cosine_precision@10": 0.09488491048593349,
|
| 83 |
+
"eval_val_evaluator_cosine_precision@5": 0.17902813299232737,
|
| 84 |
+
"eval_val_evaluator_cosine_recall@1": 0.4680306905370844,
|
| 85 |
+
"eval_val_evaluator_cosine_recall@10": 0.948849104859335,
|
| 86 |
+
"eval_val_evaluator_cosine_recall@5": 0.8951406649616368,
|
| 87 |
+
"eval_val_evaluator_dot_accuracy@1": 0.4680306905370844,
|
| 88 |
+
"eval_val_evaluator_dot_accuracy@10": 0.948849104859335,
|
| 89 |
+
"eval_val_evaluator_dot_accuracy@5": 0.8951406649616368,
|
| 90 |
+
"eval_val_evaluator_dot_map@100": 0.6501350929872222,
|
| 91 |
+
"eval_val_evaluator_dot_mrr@10": 0.6475830187147313,
|
| 92 |
+
"eval_val_evaluator_dot_mrr@100": 0.6501350929872212,
|
| 93 |
+
"eval_val_evaluator_dot_mrr@5": 0.6400468883205445,
|
| 94 |
+
"eval_val_evaluator_dot_ndcg@10": 0.7221983441504519,
|
| 95 |
+
"eval_val_evaluator_dot_ndcg@100": 0.7332422183636672,
|
| 96 |
+
"eval_val_evaluator_dot_ndcg@5": 0.7044596320615103,
|
| 97 |
+
"eval_val_evaluator_dot_precision@1": 0.4680306905370844,
|
| 98 |
+
"eval_val_evaluator_dot_precision@10": 0.09488491048593349,
|
| 99 |
+
"eval_val_evaluator_dot_precision@5": 0.17902813299232737,
|
| 100 |
+
"eval_val_evaluator_dot_recall@1": 0.4680306905370844,
|
| 101 |
+
"eval_val_evaluator_dot_recall@10": 0.948849104859335,
|
| 102 |
+
"eval_val_evaluator_dot_recall@5": 0.8951406649616368,
|
| 103 |
+
"step": 1000
|
| 104 |
+
},
|
| 105 |
+
{
|
| 106 |
+
"epoch": 6.818181818181818,
|
| 107 |
+
"grad_norm": 1.7721449136734009,
|
| 108 |
+
"learning_rate": 9.217171717171718e-06,
|
| 109 |
+
"loss": 0.0485,
|
| 110 |
+
"step": 1500
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 6.818181818181818,
|
| 114 |
+
"eval_loss": 0.06208512559533119,
|
| 115 |
+
"eval_runtime": 16.1354,
|
| 116 |
+
"eval_samples_per_second": 48.465,
|
| 117 |
+
"eval_steps_per_second": 6.074,
|
| 118 |
+
"eval_val_evaluator_cosine_accuracy@1": 0.4846547314578005,
|
| 119 |
+
"eval_val_evaluator_cosine_accuracy@10": 0.9616368286445013,
|
| 120 |
+
"eval_val_evaluator_cosine_accuracy@5": 0.921994884910486,
|
| 121 |
+
"eval_val_evaluator_cosine_map@100": 0.6677839316557986,
|
| 122 |
+
"eval_val_evaluator_cosine_mrr@10": 0.6659811025859614,
|
| 123 |
+
"eval_val_evaluator_cosine_mrr@100": 0.6677839316557982,
|
| 124 |
+
"eval_val_evaluator_cosine_mrr@5": 0.6605924978687119,
|
| 125 |
+
"eval_val_evaluator_cosine_ndcg@10": 0.739448593920066,
|
| 126 |
+
"eval_val_evaluator_cosine_ndcg@100": 0.7474718746893738,
|
| 127 |
+
"eval_val_evaluator_cosine_ndcg@5": 0.7265352392347777,
|
| 128 |
+
"eval_val_evaluator_cosine_precision@1": 0.4846547314578005,
|
| 129 |
+
"eval_val_evaluator_cosine_precision@10": 0.09616368286445011,
|
| 130 |
+
"eval_val_evaluator_cosine_precision@5": 0.18439897698209717,
|
| 131 |
+
"eval_val_evaluator_cosine_recall@1": 0.4846547314578005,
|
| 132 |
+
"eval_val_evaluator_cosine_recall@10": 0.9616368286445013,
|
| 133 |
+
"eval_val_evaluator_cosine_recall@5": 0.921994884910486,
|
| 134 |
+
"eval_val_evaluator_dot_accuracy@1": 0.4846547314578005,
|
| 135 |
+
"eval_val_evaluator_dot_accuracy@10": 0.9616368286445013,
|
| 136 |
+
"eval_val_evaluator_dot_accuracy@5": 0.921994884910486,
|
| 137 |
+
"eval_val_evaluator_dot_map@100": 0.6677839316557986,
|
| 138 |
+
"eval_val_evaluator_dot_mrr@10": 0.6659811025859614,
|
| 139 |
+
"eval_val_evaluator_dot_mrr@100": 0.6677839316557982,
|
| 140 |
+
"eval_val_evaluator_dot_mrr@5": 0.6605924978687119,
|
| 141 |
+
"eval_val_evaluator_dot_ndcg@10": 0.739448593920066,
|
| 142 |
+
"eval_val_evaluator_dot_ndcg@100": 0.7474718746893738,
|
| 143 |
+
"eval_val_evaluator_dot_ndcg@5": 0.7265352392347777,
|
| 144 |
+
"eval_val_evaluator_dot_precision@1": 0.4846547314578005,
|
| 145 |
+
"eval_val_evaluator_dot_precision@10": 0.09616368286445011,
|
| 146 |
+
"eval_val_evaluator_dot_precision@5": 0.18439897698209717,
|
| 147 |
+
"eval_val_evaluator_dot_recall@1": 0.4846547314578005,
|
| 148 |
+
"eval_val_evaluator_dot_recall@10": 0.9616368286445013,
|
| 149 |
+
"eval_val_evaluator_dot_recall@5": 0.921994884910486,
|
| 150 |
+
"step": 1500
|
| 151 |
+
},
|
| 152 |
+
{
|
| 153 |
+
"epoch": 9.090909090909092,
|
| 154 |
+
"grad_norm": 4.445181846618652,
|
| 155 |
+
"learning_rate": 8.585858585858587e-06,
|
| 156 |
+
"loss": 0.0361,
|
| 157 |
+
"step": 2000
|
| 158 |
+
},
|
| 159 |
+
{
|
| 160 |
+
"epoch": 9.090909090909092,
|
| 161 |
+
"eval_loss": 0.06154213100671768,
|
| 162 |
+
"eval_runtime": 16.1225,
|
| 163 |
+
"eval_samples_per_second": 48.504,
|
| 164 |
+
"eval_steps_per_second": 6.078,
|
| 165 |
+
"eval_val_evaluator_cosine_accuracy@1": 0.4859335038363171,
|
| 166 |
+
"eval_val_evaluator_cosine_accuracy@10": 0.9654731457800512,
|
| 167 |
+
"eval_val_evaluator_cosine_accuracy@5": 0.9271099744245525,
|
| 168 |
+
"eval_val_evaluator_cosine_map@100": 0.6706887408572961,
|
| 169 |
+
"eval_val_evaluator_cosine_mrr@10": 0.6692648073722237,
|
| 170 |
+
"eval_val_evaluator_cosine_mrr@100": 0.6706887408572959,
|
| 171 |
+
"eval_val_evaluator_cosine_mrr@5": 0.6639812446717808,
|
| 172 |
+
"eval_val_evaluator_cosine_ndcg@10": 0.7429450115476419,
|
| 173 |
+
"eval_val_evaluator_cosine_ndcg@100": 0.749838617287201,
|
| 174 |
+
"eval_val_evaluator_cosine_ndcg@5": 0.7303784552891699,
|
| 175 |
+
"eval_val_evaluator_cosine_precision@1": 0.4859335038363171,
|
| 176 |
+
"eval_val_evaluator_cosine_precision@10": 0.09654731457800511,
|
| 177 |
+
"eval_val_evaluator_cosine_precision@5": 0.18542199488491048,
|
| 178 |
+
"eval_val_evaluator_cosine_recall@1": 0.4859335038363171,
|
| 179 |
+
"eval_val_evaluator_cosine_recall@10": 0.9654731457800512,
|
| 180 |
+
"eval_val_evaluator_cosine_recall@5": 0.9271099744245525,
|
| 181 |
+
"eval_val_evaluator_dot_accuracy@1": 0.4859335038363171,
|
| 182 |
+
"eval_val_evaluator_dot_accuracy@10": 0.9654731457800512,
|
| 183 |
+
"eval_val_evaluator_dot_accuracy@5": 0.9271099744245525,
|
| 184 |
+
"eval_val_evaluator_dot_map@100": 0.6706887408572961,
|
| 185 |
+
"eval_val_evaluator_dot_mrr@10": 0.6692648073722237,
|
| 186 |
+
"eval_val_evaluator_dot_mrr@100": 0.6706887408572959,
|
| 187 |
+
"eval_val_evaluator_dot_mrr@5": 0.6639812446717808,
|
| 188 |
+
"eval_val_evaluator_dot_ndcg@10": 0.7429450115476419,
|
| 189 |
+
"eval_val_evaluator_dot_ndcg@100": 0.749838617287201,
|
| 190 |
+
"eval_val_evaluator_dot_ndcg@5": 0.7303784552891699,
|
| 191 |
+
"eval_val_evaluator_dot_precision@1": 0.4859335038363171,
|
| 192 |
+
"eval_val_evaluator_dot_precision@10": 0.09654731457800511,
|
| 193 |
+
"eval_val_evaluator_dot_precision@5": 0.18542199488491048,
|
| 194 |
+
"eval_val_evaluator_dot_recall@1": 0.4859335038363171,
|
| 195 |
+
"eval_val_evaluator_dot_recall@10": 0.9654731457800512,
|
| 196 |
+
"eval_val_evaluator_dot_recall@5": 0.9271099744245525,
|
| 197 |
+
"step": 2000
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 11.363636363636363,
|
| 201 |
+
"grad_norm": 7.720225811004639,
|
| 202 |
+
"learning_rate": 7.954545454545455e-06,
|
| 203 |
+
"loss": 0.0301,
|
| 204 |
+
"step": 2500
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 11.363636363636363,
|
| 208 |
+
"eval_loss": 0.06867850571870804,
|
| 209 |
+
"eval_runtime": 16.1559,
|
| 210 |
+
"eval_samples_per_second": 48.403,
|
| 211 |
+
"eval_steps_per_second": 6.066,
|
| 212 |
+
"eval_val_evaluator_cosine_accuracy@1": 0.49232736572890023,
|
| 213 |
+
"eval_val_evaluator_cosine_accuracy@10": 0.9603580562659847,
|
| 214 |
+
"eval_val_evaluator_cosine_accuracy@5": 0.928388746803069,
|
| 215 |
+
"eval_val_evaluator_cosine_map@100": 0.6764517431036522,
|
| 216 |
+
"eval_val_evaluator_cosine_mrr@10": 0.6745762797872769,
|
| 217 |
+
"eval_val_evaluator_cosine_mrr@100": 0.6764517431036515,
|
| 218 |
+
"eval_val_evaluator_cosine_mrr@5": 0.6698635976129574,
|
| 219 |
+
"eval_val_evaluator_cosine_ndcg@10": 0.7459669546464727,
|
| 220 |
+
"eval_val_evaluator_cosine_ndcg@100": 0.7541025758659824,
|
| 221 |
+
"eval_val_evaluator_cosine_ndcg@5": 0.7351837372969686,
|
| 222 |
+
"eval_val_evaluator_cosine_precision@1": 0.49232736572890023,
|
| 223 |
+
"eval_val_evaluator_cosine_precision@10": 0.09603580562659846,
|
| 224 |
+
"eval_val_evaluator_cosine_precision@5": 0.1856777493606138,
|
| 225 |
+
"eval_val_evaluator_cosine_recall@1": 0.49232736572890023,
|
| 226 |
+
"eval_val_evaluator_cosine_recall@10": 0.9603580562659847,
|
| 227 |
+
"eval_val_evaluator_cosine_recall@5": 0.928388746803069,
|
| 228 |
+
"eval_val_evaluator_dot_accuracy@1": 0.49232736572890023,
|
| 229 |
+
"eval_val_evaluator_dot_accuracy@10": 0.9603580562659847,
|
| 230 |
+
"eval_val_evaluator_dot_accuracy@5": 0.928388746803069,
|
| 231 |
+
"eval_val_evaluator_dot_map@100": 0.6764517431036522,
|
| 232 |
+
"eval_val_evaluator_dot_mrr@10": 0.6745762797872769,
|
| 233 |
+
"eval_val_evaluator_dot_mrr@100": 0.6764517431036515,
|
| 234 |
+
"eval_val_evaluator_dot_mrr@5": 0.6698635976129574,
|
| 235 |
+
"eval_val_evaluator_dot_ndcg@10": 0.7459669546464727,
|
| 236 |
+
"eval_val_evaluator_dot_ndcg@100": 0.7541025758659824,
|
| 237 |
+
"eval_val_evaluator_dot_ndcg@5": 0.7351837372969686,
|
| 238 |
+
"eval_val_evaluator_dot_precision@1": 0.49232736572890023,
|
| 239 |
+
"eval_val_evaluator_dot_precision@10": 0.09603580562659846,
|
| 240 |
+
"eval_val_evaluator_dot_precision@5": 0.1856777493606138,
|
| 241 |
+
"eval_val_evaluator_dot_recall@1": 0.49232736572890023,
|
| 242 |
+
"eval_val_evaluator_dot_recall@10": 0.9603580562659847,
|
| 243 |
+
"eval_val_evaluator_dot_recall@5": 0.928388746803069,
|
| 244 |
+
"step": 2500
|
| 245 |
+
},
|
| 246 |
+
{
|
| 247 |
+
"epoch": 13.636363636363637,
|
| 248 |
+
"grad_norm": 3.5265557765960693,
|
| 249 |
+
"learning_rate": 7.323232323232324e-06,
|
| 250 |
+
"loss": 0.0274,
|
| 251 |
+
"step": 3000
|
| 252 |
+
},
|
| 253 |
+
{
|
| 254 |
+
"epoch": 13.636363636363637,
|
| 255 |
+
"eval_loss": 0.06610127538442612,
|
| 256 |
+
"eval_runtime": 16.1447,
|
| 257 |
+
"eval_samples_per_second": 48.437,
|
| 258 |
+
"eval_steps_per_second": 6.07,
|
| 259 |
+
"eval_val_evaluator_cosine_accuracy@1": 0.48721227621483376,
|
| 260 |
+
"eval_val_evaluator_cosine_accuracy@10": 0.9603580562659847,
|
| 261 |
+
"eval_val_evaluator_cosine_accuracy@5": 0.9245524296675192,
|
| 262 |
+
"eval_val_evaluator_cosine_map@100": 0.6732921649074852,
|
| 263 |
+
"eval_val_evaluator_cosine_mrr@10": 0.671490987699427,
|
| 264 |
+
"eval_val_evaluator_cosine_mrr@100": 0.6732921649074848,
|
| 265 |
+
"eval_val_evaluator_cosine_mrr@5": 0.6664748508098882,
|
| 266 |
+
"eval_val_evaluator_cosine_ndcg@10": 0.7435664719550824,
|
| 267 |
+
"eval_val_evaluator_cosine_ndcg@100": 0.7519530408382193,
|
| 268 |
+
"eval_val_evaluator_cosine_ndcg@5": 0.7317557792940828,
|
| 269 |
+
"eval_val_evaluator_cosine_precision@1": 0.48721227621483376,
|
| 270 |
+
"eval_val_evaluator_cosine_precision@10": 0.09603580562659846,
|
| 271 |
+
"eval_val_evaluator_cosine_precision@5": 0.18491048593350382,
|
| 272 |
+
"eval_val_evaluator_cosine_recall@1": 0.48721227621483376,
|
| 273 |
+
"eval_val_evaluator_cosine_recall@10": 0.9603580562659847,
|
| 274 |
+
"eval_val_evaluator_cosine_recall@5": 0.9245524296675192,
|
| 275 |
+
"eval_val_evaluator_dot_accuracy@1": 0.48721227621483376,
|
| 276 |
+
"eval_val_evaluator_dot_accuracy@10": 0.9603580562659847,
|
| 277 |
+
"eval_val_evaluator_dot_accuracy@5": 0.9245524296675192,
|
| 278 |
+
"eval_val_evaluator_dot_map@100": 0.6732921649074852,
|
| 279 |
+
"eval_val_evaluator_dot_mrr@10": 0.671490987699427,
|
| 280 |
+
"eval_val_evaluator_dot_mrr@100": 0.6732921649074848,
|
| 281 |
+
"eval_val_evaluator_dot_mrr@5": 0.6664748508098882,
|
| 282 |
+
"eval_val_evaluator_dot_ndcg@10": 0.7435664719550824,
|
| 283 |
+
"eval_val_evaluator_dot_ndcg@100": 0.7519530408382193,
|
| 284 |
+
"eval_val_evaluator_dot_ndcg@5": 0.7317557792940828,
|
| 285 |
+
"eval_val_evaluator_dot_precision@1": 0.48721227621483376,
|
| 286 |
+
"eval_val_evaluator_dot_precision@10": 0.09603580562659846,
|
| 287 |
+
"eval_val_evaluator_dot_precision@5": 0.18491048593350382,
|
| 288 |
+
"eval_val_evaluator_dot_recall@1": 0.48721227621483376,
|
| 289 |
+
"eval_val_evaluator_dot_recall@10": 0.9603580562659847,
|
| 290 |
+
"eval_val_evaluator_dot_recall@5": 0.9245524296675192,
|
| 291 |
+
"step": 3000
|
| 292 |
+
},
|
| 293 |
+
{
|
| 294 |
+
"epoch": 15.909090909090908,
|
| 295 |
+
"grad_norm": 2.537912130355835,
|
| 296 |
+
"learning_rate": 6.691919191919193e-06,
|
| 297 |
+
"loss": 0.0223,
|
| 298 |
+
"step": 3500
|
| 299 |
+
},
|
| 300 |
+
{
|
| 301 |
+
"epoch": 15.909090909090908,
|
| 302 |
+
"eval_loss": 0.06057285517454147,
|
| 303 |
+
"eval_runtime": 16.2937,
|
| 304 |
+
"eval_samples_per_second": 47.994,
|
| 305 |
+
"eval_steps_per_second": 6.015,
|
| 306 |
+
"eval_val_evaluator_cosine_accuracy@1": 0.5051150895140665,
|
| 307 |
+
"eval_val_evaluator_cosine_accuracy@10": 0.959079283887468,
|
| 308 |
+
"eval_val_evaluator_cosine_accuracy@5": 0.928388746803069,
|
| 309 |
+
"eval_val_evaluator_cosine_map@100": 0.6822101368749947,
|
| 310 |
+
"eval_val_evaluator_cosine_mrr@10": 0.680220131530873,
|
| 311 |
+
"eval_val_evaluator_cosine_mrr@100": 0.6822101368749943,
|
| 312 |
+
"eval_val_evaluator_cosine_mrr@5": 0.6759164535379364,
|
| 313 |
+
"eval_val_evaluator_cosine_ndcg@10": 0.749745056809794,
|
| 314 |
+
"eval_val_evaluator_cosine_ndcg@100": 0.7586065473558236,
|
| 315 |
+
"eval_val_evaluator_cosine_ndcg@5": 0.7396139266559898,
|
| 316 |
+
"eval_val_evaluator_cosine_precision@1": 0.5051150895140665,
|
| 317 |
+
"eval_val_evaluator_cosine_precision@10": 0.0959079283887468,
|
| 318 |
+
"eval_val_evaluator_cosine_precision@5": 0.1856777493606138,
|
| 319 |
+
"eval_val_evaluator_cosine_recall@1": 0.5051150895140665,
|
| 320 |
+
"eval_val_evaluator_cosine_recall@10": 0.959079283887468,
|
| 321 |
+
"eval_val_evaluator_cosine_recall@5": 0.928388746803069,
|
| 322 |
+
"eval_val_evaluator_dot_accuracy@1": 0.5051150895140665,
|
| 323 |
+
"eval_val_evaluator_dot_accuracy@10": 0.959079283887468,
|
| 324 |
+
"eval_val_evaluator_dot_accuracy@5": 0.928388746803069,
|
| 325 |
+
"eval_val_evaluator_dot_map@100": 0.6822101368749947,
|
| 326 |
+
"eval_val_evaluator_dot_mrr@10": 0.680220131530873,
|
| 327 |
+
"eval_val_evaluator_dot_mrr@100": 0.6822101368749943,
|
| 328 |
+
"eval_val_evaluator_dot_mrr@5": 0.6759164535379364,
|
| 329 |
+
"eval_val_evaluator_dot_ndcg@10": 0.749745056809794,
|
| 330 |
+
"eval_val_evaluator_dot_ndcg@100": 0.7586065473558236,
|
| 331 |
+
"eval_val_evaluator_dot_ndcg@5": 0.7396139266559898,
|
| 332 |
+
"eval_val_evaluator_dot_precision@1": 0.5051150895140665,
|
| 333 |
+
"eval_val_evaluator_dot_precision@10": 0.0959079283887468,
|
| 334 |
+
"eval_val_evaluator_dot_precision@5": 0.1856777493606138,
|
| 335 |
+
"eval_val_evaluator_dot_recall@1": 0.5051150895140665,
|
| 336 |
+
"eval_val_evaluator_dot_recall@10": 0.959079283887468,
|
| 337 |
+
"eval_val_evaluator_dot_recall@5": 0.928388746803069,
|
| 338 |
+
"step": 3500
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 18.181818181818183,
|
| 342 |
+
"grad_norm": 6.390940189361572,
|
| 343 |
+
"learning_rate": 6.060606060606061e-06,
|
| 344 |
+
"loss": 0.021,
|
| 345 |
+
"step": 4000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 18.181818181818183,
|
| 349 |
+
"eval_loss": 0.05628308653831482,
|
| 350 |
+
"eval_runtime": 16.1509,
|
| 351 |
+
"eval_samples_per_second": 48.418,
|
| 352 |
+
"eval_steps_per_second": 6.068,
|
| 353 |
+
"eval_val_evaluator_cosine_accuracy@1": 0.5063938618925832,
|
| 354 |
+
"eval_val_evaluator_cosine_accuracy@10": 0.9552429667519181,
|
| 355 |
+
"eval_val_evaluator_cosine_accuracy@5": 0.9245524296675192,
|
| 356 |
+
"eval_val_evaluator_cosine_map@100": 0.6833838272532422,
|
| 357 |
+
"eval_val_evaluator_cosine_mrr@10": 0.6811883449031781,
|
| 358 |
+
"eval_val_evaluator_cosine_mrr@100": 0.6833838272532414,
|
| 359 |
+
"eval_val_evaluator_cosine_mrr@5": 0.6767689684569473,
|
| 360 |
+
"eval_val_evaluator_cosine_ndcg@10": 0.7496397009089436,
|
| 361 |
+
"eval_val_evaluator_cosine_ndcg@100": 0.7591798763800799,
|
| 362 |
+
"eval_val_evaluator_cosine_ndcg@5": 0.7393934605063354,
|
| 363 |
+
"eval_val_evaluator_cosine_precision@1": 0.5063938618925832,
|
| 364 |
+
"eval_val_evaluator_cosine_precision@10": 0.0955242966751918,
|
| 365 |
+
"eval_val_evaluator_cosine_precision@5": 0.1849104859335038,
|
| 366 |
+
"eval_val_evaluator_cosine_recall@1": 0.5063938618925832,
|
| 367 |
+
"eval_val_evaluator_cosine_recall@10": 0.9552429667519181,
|
| 368 |
+
"eval_val_evaluator_cosine_recall@5": 0.9245524296675192,
|
| 369 |
+
"eval_val_evaluator_dot_accuracy@1": 0.5063938618925832,
|
| 370 |
+
"eval_val_evaluator_dot_accuracy@10": 0.9552429667519181,
|
| 371 |
+
"eval_val_evaluator_dot_accuracy@5": 0.9245524296675192,
|
| 372 |
+
"eval_val_evaluator_dot_map@100": 0.6833838272532422,
|
| 373 |
+
"eval_val_evaluator_dot_mrr@10": 0.6811883449031781,
|
| 374 |
+
"eval_val_evaluator_dot_mrr@100": 0.6833838272532414,
|
| 375 |
+
"eval_val_evaluator_dot_mrr@5": 0.6767689684569473,
|
| 376 |
+
"eval_val_evaluator_dot_ndcg@10": 0.7496397009089436,
|
| 377 |
+
"eval_val_evaluator_dot_ndcg@100": 0.7591798763800799,
|
| 378 |
+
"eval_val_evaluator_dot_ndcg@5": 0.7393934605063354,
|
| 379 |
+
"eval_val_evaluator_dot_precision@1": 0.5063938618925832,
|
| 380 |
+
"eval_val_evaluator_dot_precision@10": 0.0955242966751918,
|
| 381 |
+
"eval_val_evaluator_dot_precision@5": 0.1849104859335038,
|
| 382 |
+
"eval_val_evaluator_dot_recall@1": 0.5063938618925832,
|
| 383 |
+
"eval_val_evaluator_dot_recall@10": 0.9552429667519181,
|
| 384 |
+
"eval_val_evaluator_dot_recall@5": 0.9245524296675192,
|
| 385 |
+
"step": 4000
|
| 386 |
+
},
|
| 387 |
+
{
|
| 388 |
+
"epoch": 20.454545454545453,
|
| 389 |
+
"grad_norm": 4.392763614654541,
|
| 390 |
+
"learning_rate": 5.429292929292929e-06,
|
| 391 |
+
"loss": 0.0203,
|
| 392 |
+
"step": 4500
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 20.454545454545453,
|
| 396 |
+
"eval_loss": 0.057314660400152206,
|
| 397 |
+
"eval_runtime": 16.1911,
|
| 398 |
+
"eval_samples_per_second": 48.298,
|
| 399 |
+
"eval_steps_per_second": 6.053,
|
| 400 |
+
"eval_val_evaluator_cosine_accuracy@1": 0.4782608695652174,
|
| 401 |
+
"eval_val_evaluator_cosine_accuracy@10": 0.9578005115089514,
|
| 402 |
+
"eval_val_evaluator_cosine_accuracy@5": 0.928388746803069,
|
| 403 |
+
"eval_val_evaluator_cosine_map@100": 0.6680531566193644,
|
| 404 |
+
"eval_val_evaluator_cosine_mrr@10": 0.6661059351276735,
|
| 405 |
+
"eval_val_evaluator_cosine_mrr@100": 0.6680531566193635,
|
| 406 |
+
"eval_val_evaluator_cosine_mrr@5": 0.6618925831202036,
|
| 407 |
+
"eval_val_evaluator_cosine_ndcg@10": 0.7390313134229293,
|
| 408 |
+
"eval_val_evaluator_cosine_ndcg@100": 0.7478157002187452,
|
| 409 |
+
"eval_val_evaluator_cosine_ndcg@5": 0.7292375411745403,
|
| 410 |
+
"eval_val_evaluator_cosine_precision@1": 0.4782608695652174,
|
| 411 |
+
"eval_val_evaluator_cosine_precision@10": 0.09578005115089515,
|
| 412 |
+
"eval_val_evaluator_cosine_precision@5": 0.1856777493606138,
|
| 413 |
+
"eval_val_evaluator_cosine_recall@1": 0.4782608695652174,
|
| 414 |
+
"eval_val_evaluator_cosine_recall@10": 0.9578005115089514,
|
| 415 |
+
"eval_val_evaluator_cosine_recall@5": 0.928388746803069,
|
| 416 |
+
"eval_val_evaluator_dot_accuracy@1": 0.4782608695652174,
|
| 417 |
+
"eval_val_evaluator_dot_accuracy@10": 0.9578005115089514,
|
| 418 |
+
"eval_val_evaluator_dot_accuracy@5": 0.928388746803069,
|
| 419 |
+
"eval_val_evaluator_dot_map@100": 0.6680531566193644,
|
| 420 |
+
"eval_val_evaluator_dot_mrr@10": 0.6661059351276735,
|
| 421 |
+
"eval_val_evaluator_dot_mrr@100": 0.6680531566193635,
|
| 422 |
+
"eval_val_evaluator_dot_mrr@5": 0.6618925831202036,
|
| 423 |
+
"eval_val_evaluator_dot_ndcg@10": 0.7390313134229293,
|
| 424 |
+
"eval_val_evaluator_dot_ndcg@100": 0.7478157002187452,
|
| 425 |
+
"eval_val_evaluator_dot_ndcg@5": 0.7292375411745403,
|
| 426 |
+
"eval_val_evaluator_dot_precision@1": 0.4782608695652174,
|
| 427 |
+
"eval_val_evaluator_dot_precision@10": 0.09578005115089515,
|
| 428 |
+
"eval_val_evaluator_dot_precision@5": 0.1856777493606138,
|
| 429 |
+
"eval_val_evaluator_dot_recall@1": 0.4782608695652174,
|
| 430 |
+
"eval_val_evaluator_dot_recall@10": 0.9578005115089514,
|
| 431 |
+
"eval_val_evaluator_dot_recall@5": 0.928388746803069,
|
| 432 |
+
"step": 4500
|
| 433 |
+
},
|
| 434 |
+
{
|
| 435 |
+
"epoch": 22.727272727272727,
|
| 436 |
+
"grad_norm": 4.935870170593262,
|
| 437 |
+
"learning_rate": 4.7979797979797985e-06,
|
| 438 |
+
"loss": 0.0212,
|
| 439 |
+
"step": 5000
|
| 440 |
+
},
|
| 441 |
+
{
|
| 442 |
+
"epoch": 22.727272727272727,
|
| 443 |
+
"eval_loss": 0.06370092183351517,
|
| 444 |
+
"eval_runtime": 16.1455,
|
| 445 |
+
"eval_samples_per_second": 48.435,
|
| 446 |
+
"eval_steps_per_second": 6.07,
|
| 447 |
+
"eval_val_evaluator_cosine_accuracy@1": 0.49872122762148335,
|
| 448 |
+
"eval_val_evaluator_cosine_accuracy@10": 0.9514066496163683,
|
| 449 |
+
"eval_val_evaluator_cosine_accuracy@5": 0.9258312020460358,
|
| 450 |
+
"eval_val_evaluator_cosine_map@100": 0.6769910008307545,
|
| 451 |
+
"eval_val_evaluator_cosine_mrr@10": 0.6745772946859899,
|
| 452 |
+
"eval_val_evaluator_cosine_mrr@100": 0.6769910008307539,
|
| 453 |
+
"eval_val_evaluator_cosine_mrr@5": 0.6707587382779191,
|
| 454 |
+
"eval_val_evaluator_cosine_ndcg@10": 0.743803897516351,
|
| 455 |
+
"eval_val_evaluator_cosine_ndcg@100": 0.7540671552876991,
|
| 456 |
+
"eval_val_evaluator_cosine_ndcg@5": 0.7351253519366493,
|
| 457 |
+
"eval_val_evaluator_cosine_precision@1": 0.49872122762148335,
|
| 458 |
+
"eval_val_evaluator_cosine_precision@10": 0.0951406649616368,
|
| 459 |
+
"eval_val_evaluator_cosine_precision@5": 0.18516624040920715,
|
| 460 |
+
"eval_val_evaluator_cosine_recall@1": 0.49872122762148335,
|
| 461 |
+
"eval_val_evaluator_cosine_recall@10": 0.9514066496163683,
|
| 462 |
+
"eval_val_evaluator_cosine_recall@5": 0.9258312020460358,
|
| 463 |
+
"eval_val_evaluator_dot_accuracy@1": 0.49872122762148335,
|
| 464 |
+
"eval_val_evaluator_dot_accuracy@10": 0.9514066496163683,
|
| 465 |
+
"eval_val_evaluator_dot_accuracy@5": 0.9258312020460358,
|
| 466 |
+
"eval_val_evaluator_dot_map@100": 0.6769910008307545,
|
| 467 |
+
"eval_val_evaluator_dot_mrr@10": 0.6745772946859899,
|
| 468 |
+
"eval_val_evaluator_dot_mrr@100": 0.6769910008307539,
|
| 469 |
+
"eval_val_evaluator_dot_mrr@5": 0.6707587382779191,
|
| 470 |
+
"eval_val_evaluator_dot_ndcg@10": 0.743803897516351,
|
| 471 |
+
"eval_val_evaluator_dot_ndcg@100": 0.7540671552876991,
|
| 472 |
+
"eval_val_evaluator_dot_ndcg@5": 0.7351253519366493,
|
| 473 |
+
"eval_val_evaluator_dot_precision@1": 0.49872122762148335,
|
| 474 |
+
"eval_val_evaluator_dot_precision@10": 0.0951406649616368,
|
| 475 |
+
"eval_val_evaluator_dot_precision@5": 0.18516624040920715,
|
| 476 |
+
"eval_val_evaluator_dot_recall@1": 0.49872122762148335,
|
| 477 |
+
"eval_val_evaluator_dot_recall@10": 0.9514066496163683,
|
| 478 |
+
"eval_val_evaluator_dot_recall@5": 0.9258312020460358,
|
| 479 |
+
"step": 5000
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 25.0,
|
| 483 |
+
"grad_norm": 8.184098243713379,
|
| 484 |
+
"learning_rate": 4.166666666666667e-06,
|
| 485 |
+
"loss": 0.018,
|
| 486 |
+
"step": 5500
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 25.0,
|
| 490 |
+
"eval_loss": 0.057994794100522995,
|
| 491 |
+
"eval_runtime": 16.1282,
|
| 492 |
+
"eval_samples_per_second": 48.487,
|
| 493 |
+
"eval_steps_per_second": 6.076,
|
| 494 |
+
"eval_val_evaluator_cosine_accuracy@1": 0.49744245524296676,
|
| 495 |
+
"eval_val_evaluator_cosine_accuracy@10": 0.9539641943734015,
|
| 496 |
+
"eval_val_evaluator_cosine_accuracy@5": 0.9245524296675192,
|
| 497 |
+
"eval_val_evaluator_cosine_map@100": 0.6780504881130015,
|
| 498 |
+
"eval_val_evaluator_cosine_mrr@10": 0.6757890837494416,
|
| 499 |
+
"eval_val_evaluator_cosine_mrr@100": 0.6780504881130013,
|
| 500 |
+
"eval_val_evaluator_cosine_mrr@5": 0.6714407502131281,
|
| 501 |
+
"eval_val_evaluator_cosine_ndcg@10": 0.7453518958660976,
|
| 502 |
+
"eval_val_evaluator_cosine_ndcg@100": 0.7550072905243007,
|
| 503 |
+
"eval_val_evaluator_cosine_ndcg@5": 0.7354200452387065,
|
| 504 |
+
"eval_val_evaluator_cosine_precision@1": 0.49744245524296676,
|
| 505 |
+
"eval_val_evaluator_cosine_precision@10": 0.09539641943734015,
|
| 506 |
+
"eval_val_evaluator_cosine_precision@5": 0.18491048593350382,
|
| 507 |
+
"eval_val_evaluator_cosine_recall@1": 0.49744245524296676,
|
| 508 |
+
"eval_val_evaluator_cosine_recall@10": 0.9539641943734015,
|
| 509 |
+
"eval_val_evaluator_cosine_recall@5": 0.9245524296675192,
|
| 510 |
+
"eval_val_evaluator_dot_accuracy@1": 0.49744245524296676,
|
| 511 |
+
"eval_val_evaluator_dot_accuracy@10": 0.9539641943734015,
|
| 512 |
+
"eval_val_evaluator_dot_accuracy@5": 0.9245524296675192,
|
| 513 |
+
"eval_val_evaluator_dot_map@100": 0.6780504881130015,
|
| 514 |
+
"eval_val_evaluator_dot_mrr@10": 0.6757890837494416,
|
| 515 |
+
"eval_val_evaluator_dot_mrr@100": 0.6780504881130013,
|
| 516 |
+
"eval_val_evaluator_dot_mrr@5": 0.6714407502131281,
|
| 517 |
+
"eval_val_evaluator_dot_ndcg@10": 0.7453518958660976,
|
| 518 |
+
"eval_val_evaluator_dot_ndcg@100": 0.7550072905243007,
|
| 519 |
+
"eval_val_evaluator_dot_ndcg@5": 0.7354200452387065,
|
| 520 |
+
"eval_val_evaluator_dot_precision@1": 0.49744245524296676,
|
| 521 |
+
"eval_val_evaluator_dot_precision@10": 0.09539641943734015,
|
| 522 |
+
"eval_val_evaluator_dot_precision@5": 0.18491048593350382,
|
| 523 |
+
"eval_val_evaluator_dot_recall@1": 0.49744245524296676,
|
| 524 |
+
"eval_val_evaluator_dot_recall@10": 0.9539641943734015,
|
| 525 |
+
"eval_val_evaluator_dot_recall@5": 0.9245524296675192,
|
| 526 |
+
"step": 5500
|
| 527 |
+
},
|
| 528 |
+
{
|
| 529 |
+
"epoch": 27.272727272727273,
|
| 530 |
+
"grad_norm": 0.004099330864846706,
|
| 531 |
+
"learning_rate": 3.5353535353535356e-06,
|
| 532 |
+
"loss": 0.0166,
|
| 533 |
+
"step": 6000
|
| 534 |
+
},
|
| 535 |
+
{
|
| 536 |
+
"epoch": 27.272727272727273,
|
| 537 |
+
"eval_loss": 0.056713227182626724,
|
| 538 |
+
"eval_runtime": 16.1679,
|
| 539 |
+
"eval_samples_per_second": 48.367,
|
| 540 |
+
"eval_steps_per_second": 6.061,
|
| 541 |
+
"eval_val_evaluator_cosine_accuracy@1": 0.4948849104859335,
|
| 542 |
+
"eval_val_evaluator_cosine_accuracy@10": 0.9552429667519181,
|
| 543 |
+
"eval_val_evaluator_cosine_accuracy@5": 0.9232736572890026,
|
| 544 |
+
"eval_val_evaluator_cosine_map@100": 0.6780550124960388,
|
| 545 |
+
"eval_val_evaluator_cosine_mrr@10": 0.675847440425445,
|
| 546 |
+
"eval_val_evaluator_cosine_mrr@100": 0.6780550124960384,
|
| 547 |
+
"eval_val_evaluator_cosine_mrr@5": 0.6710358056265976,
|
| 548 |
+
"eval_val_evaluator_cosine_ndcg@10": 0.7457753760186268,
|
| 549 |
+
"eval_val_evaluator_cosine_ndcg@100": 0.7551473066673557,
|
| 550 |
+
"eval_val_evaluator_cosine_ndcg@5": 0.7348943591506862,
|
| 551 |
+
"eval_val_evaluator_cosine_precision@1": 0.4948849104859335,
|
| 552 |
+
"eval_val_evaluator_cosine_precision@10": 0.0955242966751918,
|
| 553 |
+
"eval_val_evaluator_cosine_precision@5": 0.18465473145780048,
|
| 554 |
+
"eval_val_evaluator_cosine_recall@1": 0.4948849104859335,
|
| 555 |
+
"eval_val_evaluator_cosine_recall@10": 0.9552429667519181,
|
| 556 |
+
"eval_val_evaluator_cosine_recall@5": 0.9232736572890026,
|
| 557 |
+
"eval_val_evaluator_dot_accuracy@1": 0.4948849104859335,
|
| 558 |
+
"eval_val_evaluator_dot_accuracy@10": 0.9552429667519181,
|
| 559 |
+
"eval_val_evaluator_dot_accuracy@5": 0.9232736572890026,
|
| 560 |
+
"eval_val_evaluator_dot_map@100": 0.6780550124960388,
|
| 561 |
+
"eval_val_evaluator_dot_mrr@10": 0.675847440425445,
|
| 562 |
+
"eval_val_evaluator_dot_mrr@100": 0.6780550124960384,
|
| 563 |
+
"eval_val_evaluator_dot_mrr@5": 0.6710358056265976,
|
| 564 |
+
"eval_val_evaluator_dot_ndcg@10": 0.7457753760186268,
|
| 565 |
+
"eval_val_evaluator_dot_ndcg@100": 0.7551473066673557,
|
| 566 |
+
"eval_val_evaluator_dot_ndcg@5": 0.7348943591506862,
|
| 567 |
+
"eval_val_evaluator_dot_precision@1": 0.4948849104859335,
|
| 568 |
+
"eval_val_evaluator_dot_precision@10": 0.0955242966751918,
|
| 569 |
+
"eval_val_evaluator_dot_precision@5": 0.18465473145780048,
|
| 570 |
+
"eval_val_evaluator_dot_recall@1": 0.4948849104859335,
|
| 571 |
+
"eval_val_evaluator_dot_recall@10": 0.9552429667519181,
|
| 572 |
+
"eval_val_evaluator_dot_recall@5": 0.9232736572890026,
|
| 573 |
+
"step": 6000
|
| 574 |
+
},
|
| 575 |
+
{
|
| 576 |
+
"epoch": 29.545454545454547,
|
| 577 |
+
"grad_norm": 0.02236221544444561,
|
| 578 |
+
"learning_rate": 2.904040404040404e-06,
|
| 579 |
+
"loss": 0.0194,
|
| 580 |
+
"step": 6500
|
| 581 |
+
},
|
| 582 |
+
{
|
| 583 |
+
"epoch": 29.545454545454547,
|
| 584 |
+
"eval_loss": 0.05420228838920593,
|
| 585 |
+
"eval_runtime": 16.1682,
|
| 586 |
+
"eval_samples_per_second": 48.366,
|
| 587 |
+
"eval_steps_per_second": 6.061,
|
| 588 |
+
"eval_val_evaluator_cosine_accuracy@1": 0.5038363171355499,
|
| 589 |
+
"eval_val_evaluator_cosine_accuracy@10": 0.9565217391304348,
|
| 590 |
+
"eval_val_evaluator_cosine_accuracy@5": 0.9271099744245525,
|
| 591 |
+
"eval_val_evaluator_cosine_map@100": 0.6834975753216492,
|
| 592 |
+
"eval_val_evaluator_cosine_mrr@10": 0.6813415946088577,
|
| 593 |
+
"eval_val_evaluator_cosine_mrr@100": 0.6834975753216488,
|
| 594 |
+
"eval_val_evaluator_cosine_mrr@5": 0.6769181585677743,
|
| 595 |
+
"eval_val_evaluator_cosine_ndcg@10": 0.7501703504734094,
|
| 596 |
+
"eval_val_evaluator_cosine_ndcg@100": 0.759281214123215,
|
| 597 |
+
"eval_val_evaluator_cosine_ndcg@5": 0.7401646982131597,
|
| 598 |
+
"eval_val_evaluator_cosine_precision@1": 0.5038363171355499,
|
| 599 |
+
"eval_val_evaluator_cosine_precision@10": 0.09565217391304348,
|
| 600 |
+
"eval_val_evaluator_cosine_precision@5": 0.18542199488491046,
|
| 601 |
+
"eval_val_evaluator_cosine_recall@1": 0.5038363171355499,
|
| 602 |
+
"eval_val_evaluator_cosine_recall@10": 0.9565217391304348,
|
| 603 |
+
"eval_val_evaluator_cosine_recall@5": 0.9271099744245525,
|
| 604 |
+
"eval_val_evaluator_dot_accuracy@1": 0.5038363171355499,
|
| 605 |
+
"eval_val_evaluator_dot_accuracy@10": 0.9565217391304348,
|
| 606 |
+
"eval_val_evaluator_dot_accuracy@5": 0.9271099744245525,
|
| 607 |
+
"eval_val_evaluator_dot_map@100": 0.6834975753216492,
|
| 608 |
+
"eval_val_evaluator_dot_mrr@10": 0.6813415946088577,
|
| 609 |
+
"eval_val_evaluator_dot_mrr@100": 0.6834975753216488,
|
| 610 |
+
"eval_val_evaluator_dot_mrr@5": 0.6769181585677743,
|
| 611 |
+
"eval_val_evaluator_dot_ndcg@10": 0.7501703504734094,
|
| 612 |
+
"eval_val_evaluator_dot_ndcg@100": 0.759281214123215,
|
| 613 |
+
"eval_val_evaluator_dot_ndcg@5": 0.7401646982131597,
|
| 614 |
+
"eval_val_evaluator_dot_precision@1": 0.5038363171355499,
|
| 615 |
+
"eval_val_evaluator_dot_precision@10": 0.09565217391304348,
|
| 616 |
+
"eval_val_evaluator_dot_precision@5": 0.18542199488491046,
|
| 617 |
+
"eval_val_evaluator_dot_recall@1": 0.5038363171355499,
|
| 618 |
+
"eval_val_evaluator_dot_recall@10": 0.9565217391304348,
|
| 619 |
+
"eval_val_evaluator_dot_recall@5": 0.9271099744245525,
|
| 620 |
+
"step": 6500
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 31.818181818181817,
|
| 624 |
+
"grad_norm": 0.03244137763977051,
|
| 625 |
+
"learning_rate": 2.2727272727272728e-06,
|
| 626 |
+
"loss": 0.0182,
|
| 627 |
+
"step": 7000
|
| 628 |
+
},
|
| 629 |
+
{
|
| 630 |
+
"epoch": 31.818181818181817,
|
| 631 |
+
"eval_loss": 0.05468086898326874,
|
| 632 |
+
"eval_runtime": 23.5877,
|
| 633 |
+
"eval_samples_per_second": 33.153,
|
| 634 |
+
"eval_steps_per_second": 4.155,
|
| 635 |
+
"eval_val_evaluator_cosine_accuracy@1": 0.5140664961636828,
|
| 636 |
+
"eval_val_evaluator_cosine_accuracy@10": 0.9578005115089514,
|
| 637 |
+
"eval_val_evaluator_cosine_accuracy@5": 0.9296675191815856,
|
| 638 |
+
"eval_val_evaluator_cosine_map@100": 0.6896503713996865,
|
| 639 |
+
"eval_val_evaluator_cosine_mrr@10": 0.6875806844476918,
|
| 640 |
+
"eval_val_evaluator_cosine_mrr@100": 0.6896503713996863,
|
| 641 |
+
"eval_val_evaluator_cosine_mrr@5": 0.6833333333333328,
|
| 642 |
+
"eval_val_evaluator_cosine_ndcg@10": 0.7551953164170838,
|
| 643 |
+
"eval_val_evaluator_cosine_ndcg@100": 0.7640059344945531,
|
| 644 |
+
"eval_val_evaluator_cosine_ndcg@5": 0.7456070216519929,
|
| 645 |
+
"eval_val_evaluator_cosine_precision@1": 0.5140664961636828,
|
| 646 |
+
"eval_val_evaluator_cosine_precision@10": 0.09578005115089515,
|
| 647 |
+
"eval_val_evaluator_cosine_precision@5": 0.1859335038363171,
|
| 648 |
+
"eval_val_evaluator_cosine_recall@1": 0.5140664961636828,
|
| 649 |
+
"eval_val_evaluator_cosine_recall@10": 0.9578005115089514,
|
| 650 |
+
"eval_val_evaluator_cosine_recall@5": 0.9296675191815856,
|
| 651 |
+
"eval_val_evaluator_dot_accuracy@1": 0.5140664961636828,
|
| 652 |
+
"eval_val_evaluator_dot_accuracy@10": 0.9578005115089514,
|
| 653 |
+
"eval_val_evaluator_dot_accuracy@5": 0.9296675191815856,
|
| 654 |
+
"eval_val_evaluator_dot_map@100": 0.6896503713996865,
|
| 655 |
+
"eval_val_evaluator_dot_mrr@10": 0.6875806844476918,
|
| 656 |
+
"eval_val_evaluator_dot_mrr@100": 0.6896503713996863,
|
| 657 |
+
"eval_val_evaluator_dot_mrr@5": 0.6833333333333328,
|
| 658 |
+
"eval_val_evaluator_dot_ndcg@10": 0.7551953164170838,
|
| 659 |
+
"eval_val_evaluator_dot_ndcg@100": 0.7640059344945531,
|
| 660 |
+
"eval_val_evaluator_dot_ndcg@5": 0.7456070216519929,
|
| 661 |
+
"eval_val_evaluator_dot_precision@1": 0.5140664961636828,
|
| 662 |
+
"eval_val_evaluator_dot_precision@10": 0.09578005115089515,
|
| 663 |
+
"eval_val_evaluator_dot_precision@5": 0.1859335038363171,
|
| 664 |
+
"eval_val_evaluator_dot_recall@1": 0.5140664961636828,
|
| 665 |
+
"eval_val_evaluator_dot_recall@10": 0.9578005115089514,
|
| 666 |
+
"eval_val_evaluator_dot_recall@5": 0.9296675191815856,
|
| 667 |
+
"step": 7000
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 34.09090909090909,
|
| 671 |
+
"grad_norm": 0.001623653806746006,
|
| 672 |
+
"learning_rate": 1.6414141414141415e-06,
|
| 673 |
+
"loss": 0.0157,
|
| 674 |
+
"step": 7500
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 34.09090909090909,
|
| 678 |
+
"eval_loss": 0.05486880987882614,
|
| 679 |
+
"eval_runtime": 16.1005,
|
| 680 |
+
"eval_samples_per_second": 48.57,
|
| 681 |
+
"eval_steps_per_second": 6.087,
|
| 682 |
+
"eval_val_evaluator_cosine_accuracy@1": 0.5140664961636828,
|
| 683 |
+
"eval_val_evaluator_cosine_accuracy@10": 0.9565217391304348,
|
| 684 |
+
"eval_val_evaluator_cosine_accuracy@5": 0.9309462915601023,
|
| 685 |
+
"eval_val_evaluator_cosine_map@100": 0.6898639676357106,
|
| 686 |
+
"eval_val_evaluator_cosine_mrr@10": 0.6877618438679816,
|
| 687 |
+
"eval_val_evaluator_cosine_mrr@100": 0.6898639676357101,
|
| 688 |
+
"eval_val_evaluator_cosine_mrr@5": 0.6839727195225911,
|
| 689 |
+
"eval_val_evaluator_cosine_ndcg@10": 0.7550036263650965,
|
| 690 |
+
"eval_val_evaluator_cosine_ndcg@100": 0.764061133822081,
|
| 691 |
+
"eval_val_evaluator_cosine_ndcg@5": 0.7463601070791411,
|
| 692 |
+
"eval_val_evaluator_cosine_precision@1": 0.5140664961636828,
|
| 693 |
+
"eval_val_evaluator_cosine_precision@10": 0.09565217391304345,
|
| 694 |
+
"eval_val_evaluator_cosine_precision@5": 0.18618925831202046,
|
| 695 |
+
"eval_val_evaluator_cosine_recall@1": 0.5140664961636828,
|
| 696 |
+
"eval_val_evaluator_cosine_recall@10": 0.9565217391304348,
|
| 697 |
+
"eval_val_evaluator_cosine_recall@5": 0.9309462915601023,
|
| 698 |
+
"eval_val_evaluator_dot_accuracy@1": 0.5140664961636828,
|
| 699 |
+
"eval_val_evaluator_dot_accuracy@10": 0.9565217391304348,
|
| 700 |
+
"eval_val_evaluator_dot_accuracy@5": 0.9309462915601023,
|
| 701 |
+
"eval_val_evaluator_dot_map@100": 0.6898639676357106,
|
| 702 |
+
"eval_val_evaluator_dot_mrr@10": 0.6877618438679816,
|
| 703 |
+
"eval_val_evaluator_dot_mrr@100": 0.6898639676357101,
|
| 704 |
+
"eval_val_evaluator_dot_mrr@5": 0.6839727195225911,
|
| 705 |
+
"eval_val_evaluator_dot_ndcg@10": 0.7550036263650965,
|
| 706 |
+
"eval_val_evaluator_dot_ndcg@100": 0.764061133822081,
|
| 707 |
+
"eval_val_evaluator_dot_ndcg@5": 0.7463601070791411,
|
| 708 |
+
"eval_val_evaluator_dot_precision@1": 0.5140664961636828,
|
| 709 |
+
"eval_val_evaluator_dot_precision@10": 0.09565217391304345,
|
| 710 |
+
"eval_val_evaluator_dot_precision@5": 0.18618925831202046,
|
| 711 |
+
"eval_val_evaluator_dot_recall@1": 0.5140664961636828,
|
| 712 |
+
"eval_val_evaluator_dot_recall@10": 0.9565217391304348,
|
| 713 |
+
"eval_val_evaluator_dot_recall@5": 0.9309462915601023,
|
| 714 |
+
"step": 7500
|
| 715 |
+
},
|
| 716 |
+
{
|
| 717 |
+
"epoch": 36.36363636363637,
|
| 718 |
+
"grad_norm": 0.06726043671369553,
|
| 719 |
+
"learning_rate": 1.01010101010101e-06,
|
| 720 |
+
"loss": 0.016,
|
| 721 |
+
"step": 8000
|
| 722 |
+
},
|
| 723 |
+
{
|
| 724 |
+
"epoch": 36.36363636363637,
|
| 725 |
+
"eval_loss": 0.05301735922694206,
|
| 726 |
+
"eval_runtime": 16.2548,
|
| 727 |
+
"eval_samples_per_second": 48.109,
|
| 728 |
+
"eval_steps_per_second": 6.029,
|
| 729 |
+
"eval_val_evaluator_cosine_accuracy@1": 0.5089514066496164,
|
| 730 |
+
"eval_val_evaluator_cosine_accuracy@10": 0.9578005115089514,
|
| 731 |
+
"eval_val_evaluator_cosine_accuracy@5": 0.9347826086956522,
|
| 732 |
+
"eval_val_evaluator_cosine_map@100": 0.6859605033261141,
|
| 733 |
+
"eval_val_evaluator_cosine_mrr@10": 0.6839808387122961,
|
| 734 |
+
"eval_val_evaluator_cosine_mrr@100": 0.6859605033261138,
|
| 735 |
+
"eval_val_evaluator_cosine_mrr@5": 0.6806052855924973,
|
| 736 |
+
"eval_val_evaluator_cosine_ndcg@10": 0.752452574251246,
|
| 737 |
+
"eval_val_evaluator_cosine_ndcg@100": 0.7611500008339412,
|
| 738 |
+
"eval_val_evaluator_cosine_ndcg@5": 0.7447092792639648,
|
| 739 |
+
"eval_val_evaluator_cosine_precision@1": 0.5089514066496164,
|
| 740 |
+
"eval_val_evaluator_cosine_precision@10": 0.09578005115089515,
|
| 741 |
+
"eval_val_evaluator_cosine_precision@5": 0.1869565217391304,
|
| 742 |
+
"eval_val_evaluator_cosine_recall@1": 0.5089514066496164,
|
| 743 |
+
"eval_val_evaluator_cosine_recall@10": 0.9578005115089514,
|
| 744 |
+
"eval_val_evaluator_cosine_recall@5": 0.9347826086956522,
|
| 745 |
+
"eval_val_evaluator_dot_accuracy@1": 0.5089514066496164,
|
| 746 |
+
"eval_val_evaluator_dot_accuracy@10": 0.9578005115089514,
|
| 747 |
+
"eval_val_evaluator_dot_accuracy@5": 0.9347826086956522,
|
| 748 |
+
"eval_val_evaluator_dot_map@100": 0.6859605033261141,
|
| 749 |
+
"eval_val_evaluator_dot_mrr@10": 0.6839808387122961,
|
| 750 |
+
"eval_val_evaluator_dot_mrr@100": 0.6859605033261138,
|
| 751 |
+
"eval_val_evaluator_dot_mrr@5": 0.6806052855924973,
|
| 752 |
+
"eval_val_evaluator_dot_ndcg@10": 0.752452574251246,
|
| 753 |
+
"eval_val_evaluator_dot_ndcg@100": 0.7611500008339412,
|
| 754 |
+
"eval_val_evaluator_dot_ndcg@5": 0.7447092792639648,
|
| 755 |
+
"eval_val_evaluator_dot_precision@1": 0.5089514066496164,
|
| 756 |
+
"eval_val_evaluator_dot_precision@10": 0.09578005115089515,
|
| 757 |
+
"eval_val_evaluator_dot_precision@5": 0.1869565217391304,
|
| 758 |
+
"eval_val_evaluator_dot_recall@1": 0.5089514066496164,
|
| 759 |
+
"eval_val_evaluator_dot_recall@10": 0.9578005115089514,
|
| 760 |
+
"eval_val_evaluator_dot_recall@5": 0.9347826086956522,
|
| 761 |
+
"step": 8000
|
| 762 |
+
},
|
| 763 |
+
{
|
| 764 |
+
"epoch": 38.63636363636363,
|
| 765 |
+
"grad_norm": 0.05244286358356476,
|
| 766 |
+
"learning_rate": 3.787878787878788e-07,
|
| 767 |
+
"loss": 0.0142,
|
| 768 |
+
"step": 8500
|
| 769 |
+
},
|
| 770 |
+
{
|
| 771 |
+
"epoch": 38.63636363636363,
|
| 772 |
+
"eval_loss": 0.05411228537559509,
|
| 773 |
+
"eval_runtime": 16.0787,
|
| 774 |
+
"eval_samples_per_second": 48.636,
|
| 775 |
+
"eval_steps_per_second": 6.095,
|
| 776 |
+
"eval_val_evaluator_cosine_accuracy@1": 0.5127877237851662,
|
| 777 |
+
"eval_val_evaluator_cosine_accuracy@10": 0.9578005115089514,
|
| 778 |
+
"eval_val_evaluator_cosine_accuracy@5": 0.9360613810741688,
|
| 779 |
+
"eval_val_evaluator_cosine_map@100": 0.6880792251529201,
|
| 780 |
+
"eval_val_evaluator_cosine_mrr@10": 0.6861121260098237,
|
| 781 |
+
"eval_val_evaluator_cosine_mrr@100": 0.6880792251529196,
|
| 782 |
+
"eval_val_evaluator_cosine_mrr@5": 0.6829497016197776,
|
| 783 |
+
"eval_val_evaluator_cosine_ndcg@10": 0.7540621914922426,
|
| 784 |
+
"eval_val_evaluator_cosine_ndcg@100": 0.7627409192698384,
|
| 785 |
+
"eval_val_evaluator_cosine_ndcg@5": 0.7467744044168642,
|
| 786 |
+
"eval_val_evaluator_cosine_precision@1": 0.5127877237851662,
|
| 787 |
+
"eval_val_evaluator_cosine_precision@10": 0.09578005115089515,
|
| 788 |
+
"eval_val_evaluator_cosine_precision@5": 0.18721227621483372,
|
| 789 |
+
"eval_val_evaluator_cosine_recall@1": 0.5127877237851662,
|
| 790 |
+
"eval_val_evaluator_cosine_recall@10": 0.9578005115089514,
|
| 791 |
+
"eval_val_evaluator_cosine_recall@5": 0.9360613810741688,
|
| 792 |
+
"eval_val_evaluator_dot_accuracy@1": 0.5127877237851662,
|
| 793 |
+
"eval_val_evaluator_dot_accuracy@10": 0.9578005115089514,
|
| 794 |
+
"eval_val_evaluator_dot_accuracy@5": 0.9360613810741688,
|
| 795 |
+
"eval_val_evaluator_dot_map@100": 0.6880792251529201,
|
| 796 |
+
"eval_val_evaluator_dot_mrr@10": 0.6861121260098237,
|
| 797 |
+
"eval_val_evaluator_dot_mrr@100": 0.6880792251529196,
|
| 798 |
+
"eval_val_evaluator_dot_mrr@5": 0.6829497016197776,
|
| 799 |
+
"eval_val_evaluator_dot_ndcg@10": 0.7540621914922426,
|
| 800 |
+
"eval_val_evaluator_dot_ndcg@100": 0.7627409192698384,
|
| 801 |
+
"eval_val_evaluator_dot_ndcg@5": 0.7467744044168642,
|
| 802 |
+
"eval_val_evaluator_dot_precision@1": 0.5127877237851662,
|
| 803 |
+
"eval_val_evaluator_dot_precision@10": 0.09578005115089515,
|
| 804 |
+
"eval_val_evaluator_dot_precision@5": 0.18721227621483372,
|
| 805 |
+
"eval_val_evaluator_dot_recall@1": 0.5127877237851662,
|
| 806 |
+
"eval_val_evaluator_dot_recall@10": 0.9578005115089514,
|
| 807 |
+
"eval_val_evaluator_dot_recall@5": 0.9360613810741688,
|
| 808 |
+
"step": 8500
|
| 809 |
+
}
|
| 810 |
+
],
|
| 811 |
+
"logging_steps": 500,
|
| 812 |
+
"max_steps": 8800,
|
| 813 |
+
"num_input_tokens_seen": 0,
|
| 814 |
+
"num_train_epochs": 40,
|
| 815 |
+
"save_steps": 500,
|
| 816 |
+
"stateful_callbacks": {
|
| 817 |
+
"TrainerControl": {
|
| 818 |
+
"args": {
|
| 819 |
+
"should_epoch_stop": false,
|
| 820 |
+
"should_evaluate": false,
|
| 821 |
+
"should_log": false,
|
| 822 |
+
"should_save": true,
|
| 823 |
+
"should_training_stop": false
|
| 824 |
+
},
|
| 825 |
+
"attributes": {}
|
| 826 |
+
}
|
| 827 |
+
},
|
| 828 |
+
"total_flos": 0.0,
|
| 829 |
+
"train_batch_size": 8,
|
| 830 |
+
"trial_name": null,
|
| 831 |
+
"trial_params": null
|
| 832 |
+
}
|
checkpoint-8500/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:47648917f49529e1cd5c4f55390eb54738fe83ae8b04b361ce299a8b142ad8a8
|
| 3 |
+
size 5304
|
checkpoint-8500/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|