Hugging Face's logo

9wimu9
/
lfqa-mt5-large-sin-v1

Text2Text Generation
Transformers
PyTorch
mt5
Inference Endpoints
Model card Files Community
9wimu9 commited on
Commit
9e36bde
·
1 Parent(s): c7d24d3

Upload folder using huggingface_hub

Browse files
Files changed (27) hide show
  1. .gitattributes +2 -0
  2. all_results.json +13 -0
  3. checkpoint-500/config.json +32 -0
  4. checkpoint-500/generation_config.json +7 -0
  5. checkpoint-500/optimizer.pt +3 -0
  6. checkpoint-500/pytorch_model.bin +3 -0
  7. checkpoint-500/rng_state.pth +3 -0
  8. checkpoint-500/scheduler.pt +3 -0
  9. checkpoint-500/special_tokens_map.json +5 -0
  10. checkpoint-500/spiece.model +3 -0
  11. checkpoint-500/tokenizer.json +3 -0
  12. checkpoint-500/tokenizer_config.json +11 -0
  13. checkpoint-500/trainer_state.json +3016 -0
  14. checkpoint-500/training_args.bin +3 -0
  15. config.json +32 -0
  16. eval_results.json +8 -0
  17. generation_config.json +7 -0
  18. pytorch_model.bin +3 -0
  19. runs/Jun14_15-16-51_nr49739vda/events.out.tfevents.1686755936.nr49739vda.311.0 +3 -0
  20. runs/Jun14_15-16-51_nr49739vda/events.out.tfevents.1686762234.nr49739vda.311.1 +3 -0
  21. special_tokens_map.json +5 -0
  22. spiece.model +3 -0
  23. tokenizer.json +3 -0
  24. tokenizer_config.json +11 -0
  25. train_results.json +8 -0
  26. trainer_state.json +0 -0
  27. training_args.bin +3 -0
.gitattributes CHANGED
@@ -32,3 +32,5 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
32
  *.zip filter=lfs diff=lfs merge=lfs -text
33
  *.zst filter=lfs diff=lfs merge=lfs -text
34
  *tfevents* filter=lfs diff=lfs merge=lfs -text
 
 
 
32
  *.zip filter=lfs diff=lfs merge=lfs -text
33
  *.zst filter=lfs diff=lfs merge=lfs -text
34
  *tfevents* filter=lfs diff=lfs merge=lfs -text
35
+ checkpoint-500/tokenizer.json filter=lfs diff=lfs merge=lfs -text
36
+ tokenizer.json filter=lfs diff=lfs merge=lfs -text
all_results.json ADDED
@@ -0,0 +1,13 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 4.99,
3
+ "eval_loss": 2.4137141704559326,
4
+ "eval_runtime": 177.0027,
5
+ "eval_samples": 19485,
6
+ "eval_samples_per_second": 110.083,
7
+ "eval_steps_per_second": 13.763,
8
+ "train_loss": 1.6896061137063163,
9
+ "train_runtime": 6113.9003,
10
+ "train_samples": 44836,
11
+ "train_samples_per_second": 36.667,
12
+ "train_steps_per_second": 0.143
13
+ }
checkpoint-500/config.json ADDED
@@ -0,0 +1,32 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "/notebooks/seq_2_seq-odqa/trial-1/mt5-large-2",
3
+ "architectures": [
4
+ "MT5ForConditionalGeneration"
5
+ ],
6
+ "d_ff": 2816,
7
+ "d_kv": 64,
8
+ "d_model": 1024,
9
+ "decoder_start_token_id": 0,
10
+ "dense_act_fn": "gelu_new",
11
+ "dropout_rate": 0.1,
12
+ "eos_token_id": 1,
13
+ "feed_forward_proj": "gated-gelu",
14
+ "initializer_factor": 1.0,
15
+ "is_encoder_decoder": true,
16
+ "is_gated_act": true,
17
+ "layer_norm_epsilon": 1e-06,
18
+ "model_type": "mt5",
19
+ "num_decoder_layers": 24,
20
+ "num_heads": 16,
21
+ "num_layers": 24,
22
+ "output_past": true,
23
+ "pad_token_id": 0,
24
+ "relative_attention_max_distance": 128,
25
+ "relative_attention_num_buckets": 32,
26
+ "tie_word_embeddings": false,
27
+ "tokenizer_class": "T5Tokenizer",
28
+ "torch_dtype": "float32",
29
+ "transformers_version": "4.31.0.dev0",
30
+ "use_cache": true,
31
+ "vocab_size": 250112
32
+ }
checkpoint-500/generation_config.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "decoder_start_token_id": 0,
4
+ "eos_token_id": 1,
5
+ "pad_token_id": 0,
6
+ "transformers_version": "4.31.0.dev0"
7
+ }
checkpoint-500/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:37151a2ef5a921525857e7569565a4a30d551c17760dd4b33f09fc298a3e4f42
3
+ size 7516233
checkpoint-500/pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ad82d01dbfafe439440e776982553f1b2860cdaf9f52bbe010055eca2f26583d
3
+ size 4918515385
checkpoint-500/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8629ad93c79cdc75e0161802220397a1e5409e14425073a67d70ffbd81f136e1
3
+ size 14503
checkpoint-500/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bb358299dc8a6bd8690d3d2c929121a8074612b9e4d0e501265e57236f1cb0d1
3
+ size 623
checkpoint-500/special_tokens_map.json ADDED
@@ -0,0 +1,5 @@
 
 
 
 
 
 
1
+ {
2
+ "eos_token": "</s>",
3
+ "pad_token": "<pad>",
4
+ "unk_token": "<unk>"
5
+ }
checkpoint-500/spiece.model ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ef78f86560d809067d12bac6c09f19a462cb3af3f54d2b8acbba26e1433125d6
3
+ size 4309802
checkpoint-500/tokenizer.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6a2367c5e6ea9fd74eb410302c1e4a85d8cc1a017d68ae472c7facdc60d95c32
3
+ size 16315384
checkpoint-500/tokenizer_config.json ADDED
@@ -0,0 +1,11 @@
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "additional_special_tokens": null,
3
+ "clean_up_tokenization_spaces": true,
4
+ "eos_token": "</s>",
5
+ "extra_ids": 0,
6
+ "model_max_length": 1000000000000000019884624838656,
7
+ "pad_token": "<pad>",
8
+ "sp_model_kwargs": {},
9
+ "tokenizer_class": "T5Tokenizer",
10
+ "unk_token": "<unk>"
11
+ }
checkpoint-500/trainer_state.json ADDED
@@ -0,0 +1,3016 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 2.8530670470756063,
5
+ "global_step": 500,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.01,
12
+ "learning_rate": 9.98857142857143e-06,
13
+ "loss": 2.0202,
14
+ "step": 1
15
+ },
16
+ {
17
+ "epoch": 0.01,
18
+ "learning_rate": 9.977142857142858e-06,
19
+ "loss": 1.9754,
20
+ "step": 2
21
+ },
22
+ {
23
+ "epoch": 0.02,
24
+ "learning_rate": 9.965714285714286e-06,
25
+ "loss": 1.9429,
26
+ "step": 3
27
+ },
28
+ {
29
+ "epoch": 0.02,
30
+ "learning_rate": 9.954285714285715e-06,
31
+ "loss": 2.0103,
32
+ "step": 4
33
+ },
34
+ {
35
+ "epoch": 0.03,
36
+ "learning_rate": 9.942857142857145e-06,
37
+ "loss": 1.9785,
38
+ "step": 5
39
+ },
40
+ {
41
+ "epoch": 0.03,
42
+ "learning_rate": 9.931428571428571e-06,
43
+ "loss": 2.012,
44
+ "step": 6
45
+ },
46
+ {
47
+ "epoch": 0.04,
48
+ "learning_rate": 9.920000000000002e-06,
49
+ "loss": 1.9989,
50
+ "step": 7
51
+ },
52
+ {
53
+ "epoch": 0.05,
54
+ "learning_rate": 9.90857142857143e-06,
55
+ "loss": 1.9499,
56
+ "step": 8
57
+ },
58
+ {
59
+ "epoch": 0.05,
60
+ "learning_rate": 9.897142857142858e-06,
61
+ "loss": 1.9444,
62
+ "step": 9
63
+ },
64
+ {
65
+ "epoch": 0.06,
66
+ "learning_rate": 9.885714285714287e-06,
67
+ "loss": 1.9597,
68
+ "step": 10
69
+ },
70
+ {
71
+ "epoch": 0.06,
72
+ "learning_rate": 9.874285714285715e-06,
73
+ "loss": 1.9654,
74
+ "step": 11
75
+ },
76
+ {
77
+ "epoch": 0.07,
78
+ "learning_rate": 9.862857142857144e-06,
79
+ "loss": 1.9811,
80
+ "step": 12
81
+ },
82
+ {
83
+ "epoch": 0.07,
84
+ "learning_rate": 9.851428571428572e-06,
85
+ "loss": 1.9988,
86
+ "step": 13
87
+ },
88
+ {
89
+ "epoch": 0.08,
90
+ "learning_rate": 9.84e-06,
91
+ "loss": 1.9983,
92
+ "step": 14
93
+ },
94
+ {
95
+ "epoch": 0.09,
96
+ "learning_rate": 9.828571428571429e-06,
97
+ "loss": 1.9565,
98
+ "step": 15
99
+ },
100
+ {
101
+ "epoch": 0.09,
102
+ "learning_rate": 9.81714285714286e-06,
103
+ "loss": 1.876,
104
+ "step": 16
105
+ },
106
+ {
107
+ "epoch": 0.1,
108
+ "learning_rate": 9.805714285714286e-06,
109
+ "loss": 1.8638,
110
+ "step": 17
111
+ },
112
+ {
113
+ "epoch": 0.1,
114
+ "learning_rate": 9.794285714285714e-06,
115
+ "loss": 1.9153,
116
+ "step": 18
117
+ },
118
+ {
119
+ "epoch": 0.11,
120
+ "learning_rate": 9.782857142857145e-06,
121
+ "loss": 1.9132,
122
+ "step": 19
123
+ },
124
+ {
125
+ "epoch": 0.11,
126
+ "learning_rate": 9.771428571428571e-06,
127
+ "loss": 1.9551,
128
+ "step": 20
129
+ },
130
+ {
131
+ "epoch": 0.12,
132
+ "learning_rate": 9.760000000000001e-06,
133
+ "loss": 1.8265,
134
+ "step": 21
135
+ },
136
+ {
137
+ "epoch": 0.13,
138
+ "learning_rate": 9.74857142857143e-06,
139
+ "loss": 1.9253,
140
+ "step": 22
141
+ },
142
+ {
143
+ "epoch": 0.13,
144
+ "learning_rate": 9.737142857142858e-06,
145
+ "loss": 1.8201,
146
+ "step": 23
147
+ },
148
+ {
149
+ "epoch": 0.14,
150
+ "learning_rate": 9.725714285714287e-06,
151
+ "loss": 1.9096,
152
+ "step": 24
153
+ },
154
+ {
155
+ "epoch": 0.14,
156
+ "learning_rate": 9.714285714285715e-06,
157
+ "loss": 1.7755,
158
+ "step": 25
159
+ },
160
+ {
161
+ "epoch": 0.15,
162
+ "learning_rate": 9.702857142857144e-06,
163
+ "loss": 1.8151,
164
+ "step": 26
165
+ },
166
+ {
167
+ "epoch": 0.15,
168
+ "learning_rate": 9.691428571428572e-06,
169
+ "loss": 1.8359,
170
+ "step": 27
171
+ },
172
+ {
173
+ "epoch": 0.16,
174
+ "learning_rate": 9.68e-06,
175
+ "loss": 1.747,
176
+ "step": 28
177
+ },
178
+ {
179
+ "epoch": 0.17,
180
+ "learning_rate": 9.668571428571429e-06,
181
+ "loss": 1.8638,
182
+ "step": 29
183
+ },
184
+ {
185
+ "epoch": 0.17,
186
+ "learning_rate": 9.657142857142859e-06,
187
+ "loss": 1.8103,
188
+ "step": 30
189
+ },
190
+ {
191
+ "epoch": 0.18,
192
+ "learning_rate": 9.645714285714286e-06,
193
+ "loss": 1.805,
194
+ "step": 31
195
+ },
196
+ {
197
+ "epoch": 0.18,
198
+ "learning_rate": 9.634285714285714e-06,
199
+ "loss": 1.7564,
200
+ "step": 32
201
+ },
202
+ {
203
+ "epoch": 0.19,
204
+ "learning_rate": 9.622857142857144e-06,
205
+ "loss": 1.6883,
206
+ "step": 33
207
+ },
208
+ {
209
+ "epoch": 0.19,
210
+ "learning_rate": 9.611428571428573e-06,
211
+ "loss": 1.8681,
212
+ "step": 34
213
+ },
214
+ {
215
+ "epoch": 0.2,
216
+ "learning_rate": 9.600000000000001e-06,
217
+ "loss": 2.0254,
218
+ "step": 35
219
+ },
220
+ {
221
+ "epoch": 0.21,
222
+ "learning_rate": 9.58857142857143e-06,
223
+ "loss": 1.8567,
224
+ "step": 36
225
+ },
226
+ {
227
+ "epoch": 0.21,
228
+ "learning_rate": 9.577142857142858e-06,
229
+ "loss": 1.9897,
230
+ "step": 37
231
+ },
232
+ {
233
+ "epoch": 0.22,
234
+ "learning_rate": 9.565714285714287e-06,
235
+ "loss": 1.9687,
236
+ "step": 38
237
+ },
238
+ {
239
+ "epoch": 0.22,
240
+ "learning_rate": 9.554285714285715e-06,
241
+ "loss": 2.0195,
242
+ "step": 39
243
+ },
244
+ {
245
+ "epoch": 0.23,
246
+ "learning_rate": 9.542857142857143e-06,
247
+ "loss": 1.8665,
248
+ "step": 40
249
+ },
250
+ {
251
+ "epoch": 0.23,
252
+ "learning_rate": 9.531428571428572e-06,
253
+ "loss": 1.7672,
254
+ "step": 41
255
+ },
256
+ {
257
+ "epoch": 0.24,
258
+ "learning_rate": 9.52e-06,
259
+ "loss": 1.8378,
260
+ "step": 42
261
+ },
262
+ {
263
+ "epoch": 0.25,
264
+ "learning_rate": 9.508571428571429e-06,
265
+ "loss": 1.8454,
266
+ "step": 43
267
+ },
268
+ {
269
+ "epoch": 0.25,
270
+ "learning_rate": 9.497142857142859e-06,
271
+ "loss": 1.7435,
272
+ "step": 44
273
+ },
274
+ {
275
+ "epoch": 0.26,
276
+ "learning_rate": 9.485714285714287e-06,
277
+ "loss": 1.7557,
278
+ "step": 45
279
+ },
280
+ {
281
+ "epoch": 0.26,
282
+ "learning_rate": 9.474285714285714e-06,
283
+ "loss": 1.7289,
284
+ "step": 46
285
+ },
286
+ {
287
+ "epoch": 0.27,
288
+ "learning_rate": 9.462857142857144e-06,
289
+ "loss": 1.6987,
290
+ "step": 47
291
+ },
292
+ {
293
+ "epoch": 0.27,
294
+ "learning_rate": 9.451428571428573e-06,
295
+ "loss": 1.7178,
296
+ "step": 48
297
+ },
298
+ {
299
+ "epoch": 0.28,
300
+ "learning_rate": 9.440000000000001e-06,
301
+ "loss": 1.7105,
302
+ "step": 49
303
+ },
304
+ {
305
+ "epoch": 0.29,
306
+ "learning_rate": 9.42857142857143e-06,
307
+ "loss": 1.8543,
308
+ "step": 50
309
+ },
310
+ {
311
+ "epoch": 0.29,
312
+ "learning_rate": 9.417142857142858e-06,
313
+ "loss": 1.7117,
314
+ "step": 51
315
+ },
316
+ {
317
+ "epoch": 0.3,
318
+ "learning_rate": 9.405714285714286e-06,
319
+ "loss": 1.8157,
320
+ "step": 52
321
+ },
322
+ {
323
+ "epoch": 0.3,
324
+ "learning_rate": 9.394285714285715e-06,
325
+ "loss": 1.6601,
326
+ "step": 53
327
+ },
328
+ {
329
+ "epoch": 0.31,
330
+ "learning_rate": 9.382857142857143e-06,
331
+ "loss": 1.6842,
332
+ "step": 54
333
+ },
334
+ {
335
+ "epoch": 0.31,
336
+ "learning_rate": 9.371428571428572e-06,
337
+ "loss": 1.7915,
338
+ "step": 55
339
+ },
340
+ {
341
+ "epoch": 0.32,
342
+ "learning_rate": 9.360000000000002e-06,
343
+ "loss": 1.7107,
344
+ "step": 56
345
+ },
346
+ {
347
+ "epoch": 0.33,
348
+ "learning_rate": 9.348571428571429e-06,
349
+ "loss": 1.7052,
350
+ "step": 57
351
+ },
352
+ {
353
+ "epoch": 0.33,
354
+ "learning_rate": 9.337142857142859e-06,
355
+ "loss": 1.6986,
356
+ "step": 58
357
+ },
358
+ {
359
+ "epoch": 0.34,
360
+ "learning_rate": 9.325714285714287e-06,
361
+ "loss": 1.7981,
362
+ "step": 59
363
+ },
364
+ {
365
+ "epoch": 0.34,
366
+ "learning_rate": 9.314285714285714e-06,
367
+ "loss": 1.6798,
368
+ "step": 60
369
+ },
370
+ {
371
+ "epoch": 0.35,
372
+ "learning_rate": 9.302857142857144e-06,
373
+ "loss": 1.753,
374
+ "step": 61
375
+ },
376
+ {
377
+ "epoch": 0.35,
378
+ "learning_rate": 9.291428571428572e-06,
379
+ "loss": 1.6584,
380
+ "step": 62
381
+ },
382
+ {
383
+ "epoch": 0.36,
384
+ "learning_rate": 9.280000000000001e-06,
385
+ "loss": 1.711,
386
+ "step": 63
387
+ },
388
+ {
389
+ "epoch": 0.37,
390
+ "learning_rate": 9.26857142857143e-06,
391
+ "loss": 1.6625,
392
+ "step": 64
393
+ },
394
+ {
395
+ "epoch": 0.37,
396
+ "learning_rate": 9.257142857142858e-06,
397
+ "loss": 1.6937,
398
+ "step": 65
399
+ },
400
+ {
401
+ "epoch": 0.38,
402
+ "learning_rate": 9.245714285714286e-06,
403
+ "loss": 1.6583,
404
+ "step": 66
405
+ },
406
+ {
407
+ "epoch": 0.38,
408
+ "learning_rate": 9.234285714285715e-06,
409
+ "loss": 1.6619,
410
+ "step": 67
411
+ },
412
+ {
413
+ "epoch": 0.39,
414
+ "learning_rate": 9.222857142857143e-06,
415
+ "loss": 1.6572,
416
+ "step": 68
417
+ },
418
+ {
419
+ "epoch": 0.39,
420
+ "learning_rate": 9.211428571428572e-06,
421
+ "loss": 1.6537,
422
+ "step": 69
423
+ },
424
+ {
425
+ "epoch": 0.4,
426
+ "learning_rate": 9.200000000000002e-06,
427
+ "loss": 1.6308,
428
+ "step": 70
429
+ },
430
+ {
431
+ "epoch": 0.41,
432
+ "learning_rate": 9.188571428571428e-06,
433
+ "loss": 1.6518,
434
+ "step": 71
435
+ },
436
+ {
437
+ "epoch": 0.41,
438
+ "learning_rate": 9.177142857142859e-06,
439
+ "loss": 1.7237,
440
+ "step": 72
441
+ },
442
+ {
443
+ "epoch": 0.42,
444
+ "learning_rate": 9.165714285714287e-06,
445
+ "loss": 1.6625,
446
+ "step": 73
447
+ },
448
+ {
449
+ "epoch": 0.42,
450
+ "learning_rate": 9.154285714285715e-06,
451
+ "loss": 1.6398,
452
+ "step": 74
453
+ },
454
+ {
455
+ "epoch": 0.43,
456
+ "learning_rate": 9.142857142857144e-06,
457
+ "loss": 1.6435,
458
+ "step": 75
459
+ },
460
+ {
461
+ "epoch": 0.43,
462
+ "learning_rate": 9.131428571428572e-06,
463
+ "loss": 1.6497,
464
+ "step": 76
465
+ },
466
+ {
467
+ "epoch": 0.44,
468
+ "learning_rate": 9.12e-06,
469
+ "loss": 1.6057,
470
+ "step": 77
471
+ },
472
+ {
473
+ "epoch": 0.45,
474
+ "learning_rate": 9.10857142857143e-06,
475
+ "loss": 1.594,
476
+ "step": 78
477
+ },
478
+ {
479
+ "epoch": 0.45,
480
+ "learning_rate": 9.097142857142858e-06,
481
+ "loss": 1.5912,
482
+ "step": 79
483
+ },
484
+ {
485
+ "epoch": 0.46,
486
+ "learning_rate": 9.085714285714286e-06,
487
+ "loss": 1.6062,
488
+ "step": 80
489
+ },
490
+ {
491
+ "epoch": 0.46,
492
+ "learning_rate": 9.074285714285716e-06,
493
+ "loss": 1.6505,
494
+ "step": 81
495
+ },
496
+ {
497
+ "epoch": 0.47,
498
+ "learning_rate": 9.062857142857143e-06,
499
+ "loss": 1.6325,
500
+ "step": 82
501
+ },
502
+ {
503
+ "epoch": 0.47,
504
+ "learning_rate": 9.051428571428571e-06,
505
+ "loss": 1.6868,
506
+ "step": 83
507
+ },
508
+ {
509
+ "epoch": 0.48,
510
+ "learning_rate": 9.040000000000002e-06,
511
+ "loss": 1.6513,
512
+ "step": 84
513
+ },
514
+ {
515
+ "epoch": 0.49,
516
+ "learning_rate": 9.028571428571428e-06,
517
+ "loss": 1.5926,
518
+ "step": 85
519
+ },
520
+ {
521
+ "epoch": 0.49,
522
+ "learning_rate": 9.017142857142858e-06,
523
+ "loss": 1.5749,
524
+ "step": 86
525
+ },
526
+ {
527
+ "epoch": 0.5,
528
+ "learning_rate": 9.005714285714287e-06,
529
+ "loss": 1.5802,
530
+ "step": 87
531
+ },
532
+ {
533
+ "epoch": 0.5,
534
+ "learning_rate": 8.994285714285715e-06,
535
+ "loss": 1.5887,
536
+ "step": 88
537
+ },
538
+ {
539
+ "epoch": 0.51,
540
+ "learning_rate": 8.982857142857144e-06,
541
+ "loss": 1.5737,
542
+ "step": 89
543
+ },
544
+ {
545
+ "epoch": 0.51,
546
+ "learning_rate": 8.971428571428572e-06,
547
+ "loss": 1.5834,
548
+ "step": 90
549
+ },
550
+ {
551
+ "epoch": 0.52,
552
+ "learning_rate": 8.96e-06,
553
+ "loss": 1.6029,
554
+ "step": 91
555
+ },
556
+ {
557
+ "epoch": 0.52,
558
+ "learning_rate": 8.948571428571429e-06,
559
+ "loss": 1.5517,
560
+ "step": 92
561
+ },
562
+ {
563
+ "epoch": 0.53,
564
+ "learning_rate": 8.937142857142857e-06,
565
+ "loss": 1.5611,
566
+ "step": 93
567
+ },
568
+ {
569
+ "epoch": 0.54,
570
+ "learning_rate": 8.925714285714286e-06,
571
+ "loss": 1.59,
572
+ "step": 94
573
+ },
574
+ {
575
+ "epoch": 0.54,
576
+ "learning_rate": 8.914285714285716e-06,
577
+ "loss": 1.5953,
578
+ "step": 95
579
+ },
580
+ {
581
+ "epoch": 0.55,
582
+ "learning_rate": 8.902857142857143e-06,
583
+ "loss": 1.5782,
584
+ "step": 96
585
+ },
586
+ {
587
+ "epoch": 0.55,
588
+ "learning_rate": 8.891428571428571e-06,
589
+ "loss": 1.6462,
590
+ "step": 97
591
+ },
592
+ {
593
+ "epoch": 0.56,
594
+ "learning_rate": 8.880000000000001e-06,
595
+ "loss": 1.5581,
596
+ "step": 98
597
+ },
598
+ {
599
+ "epoch": 0.56,
600
+ "learning_rate": 8.86857142857143e-06,
601
+ "loss": 1.6,
602
+ "step": 99
603
+ },
604
+ {
605
+ "epoch": 0.57,
606
+ "learning_rate": 8.857142857142858e-06,
607
+ "loss": 1.5814,
608
+ "step": 100
609
+ },
610
+ {
611
+ "epoch": 0.58,
612
+ "learning_rate": 8.845714285714287e-06,
613
+ "loss": 1.5749,
614
+ "step": 101
615
+ },
616
+ {
617
+ "epoch": 0.58,
618
+ "learning_rate": 8.834285714285715e-06,
619
+ "loss": 1.5351,
620
+ "step": 102
621
+ },
622
+ {
623
+ "epoch": 0.59,
624
+ "learning_rate": 8.822857142857144e-06,
625
+ "loss": 1.56,
626
+ "step": 103
627
+ },
628
+ {
629
+ "epoch": 0.59,
630
+ "learning_rate": 8.811428571428572e-06,
631
+ "loss": 1.5687,
632
+ "step": 104
633
+ },
634
+ {
635
+ "epoch": 0.6,
636
+ "learning_rate": 8.8e-06,
637
+ "loss": 1.5763,
638
+ "step": 105
639
+ },
640
+ {
641
+ "epoch": 0.6,
642
+ "learning_rate": 8.788571428571429e-06,
643
+ "loss": 1.5875,
644
+ "step": 106
645
+ },
646
+ {
647
+ "epoch": 0.61,
648
+ "learning_rate": 8.777142857142857e-06,
649
+ "loss": 1.6139,
650
+ "step": 107
651
+ },
652
+ {
653
+ "epoch": 0.62,
654
+ "learning_rate": 8.765714285714286e-06,
655
+ "loss": 1.5734,
656
+ "step": 108
657
+ },
658
+ {
659
+ "epoch": 0.62,
660
+ "learning_rate": 8.754285714285716e-06,
661
+ "loss": 1.5117,
662
+ "step": 109
663
+ },
664
+ {
665
+ "epoch": 0.63,
666
+ "learning_rate": 8.742857142857144e-06,
667
+ "loss": 1.5611,
668
+ "step": 110
669
+ },
670
+ {
671
+ "epoch": 0.63,
672
+ "learning_rate": 8.731428571428571e-06,
673
+ "loss": 1.5481,
674
+ "step": 111
675
+ },
676
+ {
677
+ "epoch": 0.64,
678
+ "learning_rate": 8.720000000000001e-06,
679
+ "loss": 1.5578,
680
+ "step": 112
681
+ },
682
+ {
683
+ "epoch": 0.64,
684
+ "learning_rate": 8.70857142857143e-06,
685
+ "loss": 1.6075,
686
+ "step": 113
687
+ },
688
+ {
689
+ "epoch": 0.65,
690
+ "learning_rate": 8.697142857142858e-06,
691
+ "loss": 1.6112,
692
+ "step": 114
693
+ },
694
+ {
695
+ "epoch": 0.66,
696
+ "learning_rate": 8.685714285714287e-06,
697
+ "loss": 1.4582,
698
+ "step": 115
699
+ },
700
+ {
701
+ "epoch": 0.66,
702
+ "learning_rate": 8.674285714285715e-06,
703
+ "loss": 1.5832,
704
+ "step": 116
705
+ },
706
+ {
707
+ "epoch": 0.67,
708
+ "learning_rate": 8.662857142857143e-06,
709
+ "loss": 1.7716,
710
+ "step": 117
711
+ },
712
+ {
713
+ "epoch": 0.67,
714
+ "learning_rate": 8.651428571428572e-06,
715
+ "loss": 1.9526,
716
+ "step": 118
717
+ },
718
+ {
719
+ "epoch": 0.68,
720
+ "learning_rate": 8.64e-06,
721
+ "loss": 1.7239,
722
+ "step": 119
723
+ },
724
+ {
725
+ "epoch": 0.68,
726
+ "learning_rate": 8.628571428571429e-06,
727
+ "loss": 1.6897,
728
+ "step": 120
729
+ },
730
+ {
731
+ "epoch": 0.69,
732
+ "learning_rate": 8.617142857142859e-06,
733
+ "loss": 1.5891,
734
+ "step": 121
735
+ },
736
+ {
737
+ "epoch": 0.7,
738
+ "learning_rate": 8.605714285714286e-06,
739
+ "loss": 1.7339,
740
+ "step": 122
741
+ },
742
+ {
743
+ "epoch": 0.7,
744
+ "learning_rate": 8.594285714285716e-06,
745
+ "loss": 1.7173,
746
+ "step": 123
747
+ },
748
+ {
749
+ "epoch": 0.71,
750
+ "learning_rate": 8.582857142857144e-06,
751
+ "loss": 1.596,
752
+ "step": 124
753
+ },
754
+ {
755
+ "epoch": 0.71,
756
+ "learning_rate": 8.571428571428571e-06,
757
+ "loss": 1.548,
758
+ "step": 125
759
+ },
760
+ {
761
+ "epoch": 0.72,
762
+ "learning_rate": 8.560000000000001e-06,
763
+ "loss": 1.5142,
764
+ "step": 126
765
+ },
766
+ {
767
+ "epoch": 0.72,
768
+ "learning_rate": 8.54857142857143e-06,
769
+ "loss": 1.5413,
770
+ "step": 127
771
+ },
772
+ {
773
+ "epoch": 0.73,
774
+ "learning_rate": 8.537142857142858e-06,
775
+ "loss": 1.5976,
776
+ "step": 128
777
+ },
778
+ {
779
+ "epoch": 0.74,
780
+ "learning_rate": 8.525714285714286e-06,
781
+ "loss": 1.5416,
782
+ "step": 129
783
+ },
784
+ {
785
+ "epoch": 0.74,
786
+ "learning_rate": 8.514285714285715e-06,
787
+ "loss": 1.5091,
788
+ "step": 130
789
+ },
790
+ {
791
+ "epoch": 0.75,
792
+ "learning_rate": 8.502857142857143e-06,
793
+ "loss": 1.56,
794
+ "step": 131
795
+ },
796
+ {
797
+ "epoch": 0.75,
798
+ "learning_rate": 8.491428571428572e-06,
799
+ "loss": 1.5188,
800
+ "step": 132
801
+ },
802
+ {
803
+ "epoch": 0.76,
804
+ "learning_rate": 8.48e-06,
805
+ "loss": 1.5378,
806
+ "step": 133
807
+ },
808
+ {
809
+ "epoch": 0.76,
810
+ "learning_rate": 8.468571428571429e-06,
811
+ "loss": 1.5595,
812
+ "step": 134
813
+ },
814
+ {
815
+ "epoch": 0.77,
816
+ "learning_rate": 8.457142857142859e-06,
817
+ "loss": 1.5591,
818
+ "step": 135
819
+ },
820
+ {
821
+ "epoch": 0.78,
822
+ "learning_rate": 8.445714285714285e-06,
823
+ "loss": 1.565,
824
+ "step": 136
825
+ },
826
+ {
827
+ "epoch": 0.78,
828
+ "learning_rate": 8.434285714285716e-06,
829
+ "loss": 1.5318,
830
+ "step": 137
831
+ },
832
+ {
833
+ "epoch": 0.79,
834
+ "learning_rate": 8.422857142857144e-06,
835
+ "loss": 1.6051,
836
+ "step": 138
837
+ },
838
+ {
839
+ "epoch": 0.79,
840
+ "learning_rate": 8.411428571428572e-06,
841
+ "loss": 1.6247,
842
+ "step": 139
843
+ },
844
+ {
845
+ "epoch": 0.8,
846
+ "learning_rate": 8.400000000000001e-06,
847
+ "loss": 1.5086,
848
+ "step": 140
849
+ },
850
+ {
851
+ "epoch": 0.8,
852
+ "learning_rate": 8.38857142857143e-06,
853
+ "loss": 1.5235,
854
+ "step": 141
855
+ },
856
+ {
857
+ "epoch": 0.81,
858
+ "learning_rate": 8.377142857142858e-06,
859
+ "loss": 1.5342,
860
+ "step": 142
861
+ },
862
+ {
863
+ "epoch": 0.82,
864
+ "learning_rate": 8.365714285714286e-06,
865
+ "loss": 1.4861,
866
+ "step": 143
867
+ },
868
+ {
869
+ "epoch": 0.82,
870
+ "learning_rate": 8.354285714285715e-06,
871
+ "loss": 1.4781,
872
+ "step": 144
873
+ },
874
+ {
875
+ "epoch": 0.83,
876
+ "learning_rate": 8.342857142857143e-06,
877
+ "loss": 1.5155,
878
+ "step": 145
879
+ },
880
+ {
881
+ "epoch": 0.83,
882
+ "learning_rate": 8.331428571428573e-06,
883
+ "loss": 1.513,
884
+ "step": 146
885
+ },
886
+ {
887
+ "epoch": 0.84,
888
+ "learning_rate": 8.32e-06,
889
+ "loss": 1.4958,
890
+ "step": 147
891
+ },
892
+ {
893
+ "epoch": 0.84,
894
+ "learning_rate": 8.308571428571428e-06,
895
+ "loss": 1.4728,
896
+ "step": 148
897
+ },
898
+ {
899
+ "epoch": 0.85,
900
+ "learning_rate": 8.297142857142859e-06,
901
+ "loss": 1.504,
902
+ "step": 149
903
+ },
904
+ {
905
+ "epoch": 0.86,
906
+ "learning_rate": 8.285714285714287e-06,
907
+ "loss": 1.5638,
908
+ "step": 150
909
+ },
910
+ {
911
+ "epoch": 0.86,
912
+ "learning_rate": 8.274285714285715e-06,
913
+ "loss": 1.4473,
914
+ "step": 151
915
+ },
916
+ {
917
+ "epoch": 0.87,
918
+ "learning_rate": 8.262857142857144e-06,
919
+ "loss": 1.4839,
920
+ "step": 152
921
+ },
922
+ {
923
+ "epoch": 0.87,
924
+ "learning_rate": 8.251428571428572e-06,
925
+ "loss": 1.5174,
926
+ "step": 153
927
+ },
928
+ {
929
+ "epoch": 0.88,
930
+ "learning_rate": 8.24e-06,
931
+ "loss": 1.5251,
932
+ "step": 154
933
+ },
934
+ {
935
+ "epoch": 0.88,
936
+ "learning_rate": 8.22857142857143e-06,
937
+ "loss": 1.4409,
938
+ "step": 155
939
+ },
940
+ {
941
+ "epoch": 0.89,
942
+ "learning_rate": 8.217142857142858e-06,
943
+ "loss": 1.469,
944
+ "step": 156
945
+ },
946
+ {
947
+ "epoch": 0.9,
948
+ "learning_rate": 8.205714285714286e-06,
949
+ "loss": 1.4872,
950
+ "step": 157
951
+ },
952
+ {
953
+ "epoch": 0.9,
954
+ "learning_rate": 8.194285714285714e-06,
955
+ "loss": 1.4819,
956
+ "step": 158
957
+ },
958
+ {
959
+ "epoch": 0.91,
960
+ "learning_rate": 8.182857142857143e-06,
961
+ "loss": 1.4822,
962
+ "step": 159
963
+ },
964
+ {
965
+ "epoch": 0.91,
966
+ "learning_rate": 8.171428571428573e-06,
967
+ "loss": 1.4797,
968
+ "step": 160
969
+ },
970
+ {
971
+ "epoch": 0.92,
972
+ "learning_rate": 8.16e-06,
973
+ "loss": 1.4872,
974
+ "step": 161
975
+ },
976
+ {
977
+ "epoch": 0.92,
978
+ "learning_rate": 8.148571428571428e-06,
979
+ "loss": 1.4472,
980
+ "step": 162
981
+ },
982
+ {
983
+ "epoch": 0.93,
984
+ "learning_rate": 8.137142857142858e-06,
985
+ "loss": 1.4962,
986
+ "step": 163
987
+ },
988
+ {
989
+ "epoch": 0.94,
990
+ "learning_rate": 8.125714285714287e-06,
991
+ "loss": 1.5026,
992
+ "step": 164
993
+ },
994
+ {
995
+ "epoch": 0.94,
996
+ "learning_rate": 8.114285714285715e-06,
997
+ "loss": 1.5028,
998
+ "step": 165
999
+ },
1000
+ {
1001
+ "epoch": 0.95,
1002
+ "learning_rate": 8.102857142857144e-06,
1003
+ "loss": 1.5236,
1004
+ "step": 166
1005
+ },
1006
+ {
1007
+ "epoch": 0.95,
1008
+ "learning_rate": 8.091428571428572e-06,
1009
+ "loss": 1.4764,
1010
+ "step": 167
1011
+ },
1012
+ {
1013
+ "epoch": 0.96,
1014
+ "learning_rate": 8.08e-06,
1015
+ "loss": 1.5006,
1016
+ "step": 168
1017
+ },
1018
+ {
1019
+ "epoch": 0.96,
1020
+ "learning_rate": 8.068571428571429e-06,
1021
+ "loss": 1.4522,
1022
+ "step": 169
1023
+ },
1024
+ {
1025
+ "epoch": 0.97,
1026
+ "learning_rate": 8.057142857142857e-06,
1027
+ "loss": 1.4775,
1028
+ "step": 170
1029
+ },
1030
+ {
1031
+ "epoch": 0.98,
1032
+ "learning_rate": 8.045714285714286e-06,
1033
+ "loss": 1.4455,
1034
+ "step": 171
1035
+ },
1036
+ {
1037
+ "epoch": 0.98,
1038
+ "learning_rate": 8.034285714285714e-06,
1039
+ "loss": 1.4824,
1040
+ "step": 172
1041
+ },
1042
+ {
1043
+ "epoch": 0.99,
1044
+ "learning_rate": 8.022857142857143e-06,
1045
+ "loss": 1.4554,
1046
+ "step": 173
1047
+ },
1048
+ {
1049
+ "epoch": 0.99,
1050
+ "learning_rate": 8.011428571428573e-06,
1051
+ "loss": 1.4831,
1052
+ "step": 174
1053
+ },
1054
+ {
1055
+ "epoch": 1.0,
1056
+ "learning_rate": 8.000000000000001e-06,
1057
+ "loss": 1.4786,
1058
+ "step": 175
1059
+ },
1060
+ {
1061
+ "epoch": 1.0,
1062
+ "learning_rate": 7.988571428571428e-06,
1063
+ "loss": 1.9457,
1064
+ "step": 176
1065
+ },
1066
+ {
1067
+ "epoch": 1.01,
1068
+ "learning_rate": 7.977142857142858e-06,
1069
+ "loss": 1.8681,
1070
+ "step": 177
1071
+ },
1072
+ {
1073
+ "epoch": 1.02,
1074
+ "learning_rate": 7.965714285714287e-06,
1075
+ "loss": 1.8721,
1076
+ "step": 178
1077
+ },
1078
+ {
1079
+ "epoch": 1.02,
1080
+ "learning_rate": 7.954285714285715e-06,
1081
+ "loss": 1.8807,
1082
+ "step": 179
1083
+ },
1084
+ {
1085
+ "epoch": 1.03,
1086
+ "learning_rate": 7.942857142857144e-06,
1087
+ "loss": 1.8907,
1088
+ "step": 180
1089
+ },
1090
+ {
1091
+ "epoch": 1.03,
1092
+ "learning_rate": 7.931428571428572e-06,
1093
+ "loss": 1.868,
1094
+ "step": 181
1095
+ },
1096
+ {
1097
+ "epoch": 1.04,
1098
+ "learning_rate": 7.92e-06,
1099
+ "loss": 1.9417,
1100
+ "step": 182
1101
+ },
1102
+ {
1103
+ "epoch": 1.04,
1104
+ "learning_rate": 7.908571428571429e-06,
1105
+ "loss": 1.8749,
1106
+ "step": 183
1107
+ },
1108
+ {
1109
+ "epoch": 1.05,
1110
+ "learning_rate": 7.897142857142857e-06,
1111
+ "loss": 1.8543,
1112
+ "step": 184
1113
+ },
1114
+ {
1115
+ "epoch": 1.06,
1116
+ "learning_rate": 7.885714285714286e-06,
1117
+ "loss": 1.9108,
1118
+ "step": 185
1119
+ },
1120
+ {
1121
+ "epoch": 1.06,
1122
+ "learning_rate": 7.874285714285716e-06,
1123
+ "loss": 1.8662,
1124
+ "step": 186
1125
+ },
1126
+ {
1127
+ "epoch": 1.07,
1128
+ "learning_rate": 7.862857142857143e-06,
1129
+ "loss": 1.8439,
1130
+ "step": 187
1131
+ },
1132
+ {
1133
+ "epoch": 1.07,
1134
+ "learning_rate": 7.851428571428573e-06,
1135
+ "loss": 1.8629,
1136
+ "step": 188
1137
+ },
1138
+ {
1139
+ "epoch": 1.08,
1140
+ "learning_rate": 7.840000000000001e-06,
1141
+ "loss": 1.8611,
1142
+ "step": 189
1143
+ },
1144
+ {
1145
+ "epoch": 1.08,
1146
+ "learning_rate": 7.828571428571428e-06,
1147
+ "loss": 1.9209,
1148
+ "step": 190
1149
+ },
1150
+ {
1151
+ "epoch": 1.09,
1152
+ "learning_rate": 7.817142857142858e-06,
1153
+ "loss": 1.8903,
1154
+ "step": 191
1155
+ },
1156
+ {
1157
+ "epoch": 1.1,
1158
+ "learning_rate": 7.805714285714286e-06,
1159
+ "loss": 1.8397,
1160
+ "step": 192
1161
+ },
1162
+ {
1163
+ "epoch": 1.1,
1164
+ "learning_rate": 7.794285714285715e-06,
1165
+ "loss": 1.8394,
1166
+ "step": 193
1167
+ },
1168
+ {
1169
+ "epoch": 1.11,
1170
+ "learning_rate": 7.782857142857143e-06,
1171
+ "loss": 1.8822,
1172
+ "step": 194
1173
+ },
1174
+ {
1175
+ "epoch": 1.11,
1176
+ "learning_rate": 7.771428571428572e-06,
1177
+ "loss": 1.8179,
1178
+ "step": 195
1179
+ },
1180
+ {
1181
+ "epoch": 1.12,
1182
+ "learning_rate": 7.76e-06,
1183
+ "loss": 1.8568,
1184
+ "step": 196
1185
+ },
1186
+ {
1187
+ "epoch": 1.12,
1188
+ "learning_rate": 7.74857142857143e-06,
1189
+ "loss": 1.8766,
1190
+ "step": 197
1191
+ },
1192
+ {
1193
+ "epoch": 1.13,
1194
+ "learning_rate": 7.737142857142857e-06,
1195
+ "loss": 1.8984,
1196
+ "step": 198
1197
+ },
1198
+ {
1199
+ "epoch": 1.14,
1200
+ "learning_rate": 7.725714285714286e-06,
1201
+ "loss": 1.9151,
1202
+ "step": 199
1203
+ },
1204
+ {
1205
+ "epoch": 1.14,
1206
+ "learning_rate": 7.714285714285716e-06,
1207
+ "loss": 1.8517,
1208
+ "step": 200
1209
+ },
1210
+ {
1211
+ "epoch": 1.15,
1212
+ "learning_rate": 7.702857142857142e-06,
1213
+ "loss": 1.8674,
1214
+ "step": 201
1215
+ },
1216
+ {
1217
+ "epoch": 1.15,
1218
+ "learning_rate": 7.691428571428573e-06,
1219
+ "loss": 1.8796,
1220
+ "step": 202
1221
+ },
1222
+ {
1223
+ "epoch": 1.16,
1224
+ "learning_rate": 7.680000000000001e-06,
1225
+ "loss": 1.9014,
1226
+ "step": 203
1227
+ },
1228
+ {
1229
+ "epoch": 1.16,
1230
+ "learning_rate": 7.66857142857143e-06,
1231
+ "loss": 1.8437,
1232
+ "step": 204
1233
+ },
1234
+ {
1235
+ "epoch": 1.17,
1236
+ "learning_rate": 7.657142857142858e-06,
1237
+ "loss": 1.8858,
1238
+ "step": 205
1239
+ },
1240
+ {
1241
+ "epoch": 1.18,
1242
+ "learning_rate": 7.645714285714286e-06,
1243
+ "loss": 1.882,
1244
+ "step": 206
1245
+ },
1246
+ {
1247
+ "epoch": 1.18,
1248
+ "learning_rate": 7.634285714285715e-06,
1249
+ "loss": 1.8613,
1250
+ "step": 207
1251
+ },
1252
+ {
1253
+ "epoch": 1.19,
1254
+ "learning_rate": 7.622857142857143e-06,
1255
+ "loss": 1.907,
1256
+ "step": 208
1257
+ },
1258
+ {
1259
+ "epoch": 1.19,
1260
+ "learning_rate": 7.611428571428572e-06,
1261
+ "loss": 1.9055,
1262
+ "step": 209
1263
+ },
1264
+ {
1265
+ "epoch": 1.2,
1266
+ "learning_rate": 7.600000000000001e-06,
1267
+ "loss": 1.8942,
1268
+ "step": 210
1269
+ },
1270
+ {
1271
+ "epoch": 1.2,
1272
+ "learning_rate": 7.588571428571429e-06,
1273
+ "loss": 1.8409,
1274
+ "step": 211
1275
+ },
1276
+ {
1277
+ "epoch": 1.21,
1278
+ "learning_rate": 7.577142857142857e-06,
1279
+ "loss": 1.826,
1280
+ "step": 212
1281
+ },
1282
+ {
1283
+ "epoch": 1.22,
1284
+ "learning_rate": 7.565714285714286e-06,
1285
+ "loss": 1.8701,
1286
+ "step": 213
1287
+ },
1288
+ {
1289
+ "epoch": 1.22,
1290
+ "learning_rate": 7.5542857142857155e-06,
1291
+ "loss": 1.8298,
1292
+ "step": 214
1293
+ },
1294
+ {
1295
+ "epoch": 1.23,
1296
+ "learning_rate": 7.542857142857144e-06,
1297
+ "loss": 1.9199,
1298
+ "step": 215
1299
+ },
1300
+ {
1301
+ "epoch": 1.23,
1302
+ "learning_rate": 7.5314285714285716e-06,
1303
+ "loss": 1.8758,
1304
+ "step": 216
1305
+ },
1306
+ {
1307
+ "epoch": 1.24,
1308
+ "learning_rate": 7.520000000000001e-06,
1309
+ "loss": 1.8709,
1310
+ "step": 217
1311
+ },
1312
+ {
1313
+ "epoch": 1.24,
1314
+ "learning_rate": 7.508571428571429e-06,
1315
+ "loss": 1.875,
1316
+ "step": 218
1317
+ },
1318
+ {
1319
+ "epoch": 1.25,
1320
+ "learning_rate": 7.497142857142857e-06,
1321
+ "loss": 1.9255,
1322
+ "step": 219
1323
+ },
1324
+ {
1325
+ "epoch": 1.26,
1326
+ "learning_rate": 7.485714285714286e-06,
1327
+ "loss": 1.8664,
1328
+ "step": 220
1329
+ },
1330
+ {
1331
+ "epoch": 1.26,
1332
+ "learning_rate": 7.4742857142857154e-06,
1333
+ "loss": 1.8337,
1334
+ "step": 221
1335
+ },
1336
+ {
1337
+ "epoch": 1.27,
1338
+ "learning_rate": 7.462857142857144e-06,
1339
+ "loss": 1.9226,
1340
+ "step": 222
1341
+ },
1342
+ {
1343
+ "epoch": 1.27,
1344
+ "learning_rate": 7.4514285714285715e-06,
1345
+ "loss": 1.8908,
1346
+ "step": 223
1347
+ },
1348
+ {
1349
+ "epoch": 1.28,
1350
+ "learning_rate": 7.440000000000001e-06,
1351
+ "loss": 1.8781,
1352
+ "step": 224
1353
+ },
1354
+ {
1355
+ "epoch": 1.28,
1356
+ "learning_rate": 7.428571428571429e-06,
1357
+ "loss": 1.9171,
1358
+ "step": 225
1359
+ },
1360
+ {
1361
+ "epoch": 1.29,
1362
+ "learning_rate": 7.417142857142857e-06,
1363
+ "loss": 1.8577,
1364
+ "step": 226
1365
+ },
1366
+ {
1367
+ "epoch": 1.3,
1368
+ "learning_rate": 7.405714285714286e-06,
1369
+ "loss": 1.8957,
1370
+ "step": 227
1371
+ },
1372
+ {
1373
+ "epoch": 1.3,
1374
+ "learning_rate": 7.394285714285715e-06,
1375
+ "loss": 1.9199,
1376
+ "step": 228
1377
+ },
1378
+ {
1379
+ "epoch": 1.31,
1380
+ "learning_rate": 7.382857142857144e-06,
1381
+ "loss": 1.949,
1382
+ "step": 229
1383
+ },
1384
+ {
1385
+ "epoch": 1.31,
1386
+ "learning_rate": 7.371428571428571e-06,
1387
+ "loss": 1.9293,
1388
+ "step": 230
1389
+ },
1390
+ {
1391
+ "epoch": 1.32,
1392
+ "learning_rate": 7.360000000000001e-06,
1393
+ "loss": 1.9471,
1394
+ "step": 231
1395
+ },
1396
+ {
1397
+ "epoch": 1.32,
1398
+ "learning_rate": 7.348571428571429e-06,
1399
+ "loss": 1.8477,
1400
+ "step": 232
1401
+ },
1402
+ {
1403
+ "epoch": 1.33,
1404
+ "learning_rate": 7.337142857142858e-06,
1405
+ "loss": 1.8893,
1406
+ "step": 233
1407
+ },
1408
+ {
1409
+ "epoch": 1.34,
1410
+ "learning_rate": 7.325714285714286e-06,
1411
+ "loss": 1.8764,
1412
+ "step": 234
1413
+ },
1414
+ {
1415
+ "epoch": 1.34,
1416
+ "learning_rate": 7.314285714285715e-06,
1417
+ "loss": 1.9136,
1418
+ "step": 235
1419
+ },
1420
+ {
1421
+ "epoch": 1.35,
1422
+ "learning_rate": 7.302857142857144e-06,
1423
+ "loss": 1.9147,
1424
+ "step": 236
1425
+ },
1426
+ {
1427
+ "epoch": 1.35,
1428
+ "learning_rate": 7.291428571428571e-06,
1429
+ "loss": 1.8703,
1430
+ "step": 237
1431
+ },
1432
+ {
1433
+ "epoch": 1.36,
1434
+ "learning_rate": 7.280000000000001e-06,
1435
+ "loss": 1.9123,
1436
+ "step": 238
1437
+ },
1438
+ {
1439
+ "epoch": 1.36,
1440
+ "learning_rate": 7.268571428571429e-06,
1441
+ "loss": 1.9115,
1442
+ "step": 239
1443
+ },
1444
+ {
1445
+ "epoch": 1.37,
1446
+ "learning_rate": 7.257142857142858e-06,
1447
+ "loss": 1.9358,
1448
+ "step": 240
1449
+ },
1450
+ {
1451
+ "epoch": 1.38,
1452
+ "learning_rate": 7.245714285714286e-06,
1453
+ "loss": 1.9283,
1454
+ "step": 241
1455
+ },
1456
+ {
1457
+ "epoch": 1.38,
1458
+ "learning_rate": 7.234285714285715e-06,
1459
+ "loss": 1.8585,
1460
+ "step": 242
1461
+ },
1462
+ {
1463
+ "epoch": 1.39,
1464
+ "learning_rate": 7.222857142857144e-06,
1465
+ "loss": 1.8692,
1466
+ "step": 243
1467
+ },
1468
+ {
1469
+ "epoch": 1.39,
1470
+ "learning_rate": 7.211428571428573e-06,
1471
+ "loss": 1.8875,
1472
+ "step": 244
1473
+ },
1474
+ {
1475
+ "epoch": 1.4,
1476
+ "learning_rate": 7.2000000000000005e-06,
1477
+ "loss": 1.9205,
1478
+ "step": 245
1479
+ },
1480
+ {
1481
+ "epoch": 1.4,
1482
+ "learning_rate": 7.188571428571429e-06,
1483
+ "loss": 1.9152,
1484
+ "step": 246
1485
+ },
1486
+ {
1487
+ "epoch": 1.41,
1488
+ "learning_rate": 7.177142857142858e-06,
1489
+ "loss": 1.8396,
1490
+ "step": 247
1491
+ },
1492
+ {
1493
+ "epoch": 1.42,
1494
+ "learning_rate": 7.165714285714286e-06,
1495
+ "loss": 1.8779,
1496
+ "step": 248
1497
+ },
1498
+ {
1499
+ "epoch": 1.42,
1500
+ "learning_rate": 7.154285714285715e-06,
1501
+ "loss": 1.8801,
1502
+ "step": 249
1503
+ },
1504
+ {
1505
+ "epoch": 1.43,
1506
+ "learning_rate": 7.1428571428571436e-06,
1507
+ "loss": 1.8648,
1508
+ "step": 250
1509
+ },
1510
+ {
1511
+ "epoch": 1.43,
1512
+ "learning_rate": 7.131428571428573e-06,
1513
+ "loss": 1.8415,
1514
+ "step": 251
1515
+ },
1516
+ {
1517
+ "epoch": 1.44,
1518
+ "learning_rate": 7.1200000000000004e-06,
1519
+ "loss": 1.9185,
1520
+ "step": 252
1521
+ },
1522
+ {
1523
+ "epoch": 1.44,
1524
+ "learning_rate": 7.108571428571429e-06,
1525
+ "loss": 1.8344,
1526
+ "step": 253
1527
+ },
1528
+ {
1529
+ "epoch": 1.45,
1530
+ "learning_rate": 7.097142857142858e-06,
1531
+ "loss": 1.8843,
1532
+ "step": 254
1533
+ },
1534
+ {
1535
+ "epoch": 1.46,
1536
+ "learning_rate": 7.085714285714286e-06,
1537
+ "loss": 1.8984,
1538
+ "step": 255
1539
+ },
1540
+ {
1541
+ "epoch": 1.46,
1542
+ "learning_rate": 7.074285714285715e-06,
1543
+ "loss": 1.8439,
1544
+ "step": 256
1545
+ },
1546
+ {
1547
+ "epoch": 1.47,
1548
+ "learning_rate": 7.0628571428571435e-06,
1549
+ "loss": 1.8614,
1550
+ "step": 257
1551
+ },
1552
+ {
1553
+ "epoch": 1.47,
1554
+ "learning_rate": 7.051428571428573e-06,
1555
+ "loss": 1.8259,
1556
+ "step": 258
1557
+ },
1558
+ {
1559
+ "epoch": 1.48,
1560
+ "learning_rate": 7.04e-06,
1561
+ "loss": 1.8883,
1562
+ "step": 259
1563
+ },
1564
+ {
1565
+ "epoch": 1.48,
1566
+ "learning_rate": 7.028571428571429e-06,
1567
+ "loss": 1.9634,
1568
+ "step": 260
1569
+ },
1570
+ {
1571
+ "epoch": 1.49,
1572
+ "learning_rate": 7.017142857142858e-06,
1573
+ "loss": 1.9071,
1574
+ "step": 261
1575
+ },
1576
+ {
1577
+ "epoch": 1.5,
1578
+ "learning_rate": 7.0057142857142865e-06,
1579
+ "loss": 1.9218,
1580
+ "step": 262
1581
+ },
1582
+ {
1583
+ "epoch": 1.5,
1584
+ "learning_rate": 6.994285714285715e-06,
1585
+ "loss": 1.8727,
1586
+ "step": 263
1587
+ },
1588
+ {
1589
+ "epoch": 1.51,
1590
+ "learning_rate": 6.982857142857143e-06,
1591
+ "loss": 1.8501,
1592
+ "step": 264
1593
+ },
1594
+ {
1595
+ "epoch": 1.51,
1596
+ "learning_rate": 6.971428571428573e-06,
1597
+ "loss": 1.8371,
1598
+ "step": 265
1599
+ },
1600
+ {
1601
+ "epoch": 1.52,
1602
+ "learning_rate": 6.96e-06,
1603
+ "loss": 1.8941,
1604
+ "step": 266
1605
+ },
1606
+ {
1607
+ "epoch": 1.52,
1608
+ "learning_rate": 6.948571428571429e-06,
1609
+ "loss": 1.8408,
1610
+ "step": 267
1611
+ },
1612
+ {
1613
+ "epoch": 1.53,
1614
+ "learning_rate": 6.937142857142858e-06,
1615
+ "loss": 1.9015,
1616
+ "step": 268
1617
+ },
1618
+ {
1619
+ "epoch": 1.53,
1620
+ "learning_rate": 6.9257142857142864e-06,
1621
+ "loss": 1.94,
1622
+ "step": 269
1623
+ },
1624
+ {
1625
+ "epoch": 1.54,
1626
+ "learning_rate": 6.914285714285715e-06,
1627
+ "loss": 1.8656,
1628
+ "step": 270
1629
+ },
1630
+ {
1631
+ "epoch": 1.55,
1632
+ "learning_rate": 6.902857142857143e-06,
1633
+ "loss": 1.8429,
1634
+ "step": 271
1635
+ },
1636
+ {
1637
+ "epoch": 1.55,
1638
+ "learning_rate": 6.891428571428573e-06,
1639
+ "loss": 1.9209,
1640
+ "step": 272
1641
+ },
1642
+ {
1643
+ "epoch": 1.56,
1644
+ "learning_rate": 6.88e-06,
1645
+ "loss": 1.9166,
1646
+ "step": 273
1647
+ },
1648
+ {
1649
+ "epoch": 1.56,
1650
+ "learning_rate": 6.868571428571429e-06,
1651
+ "loss": 1.9284,
1652
+ "step": 274
1653
+ },
1654
+ {
1655
+ "epoch": 1.57,
1656
+ "learning_rate": 6.857142857142858e-06,
1657
+ "loss": 1.8919,
1658
+ "step": 275
1659
+ },
1660
+ {
1661
+ "epoch": 1.57,
1662
+ "learning_rate": 6.845714285714286e-06,
1663
+ "loss": 1.9217,
1664
+ "step": 276
1665
+ },
1666
+ {
1667
+ "epoch": 1.58,
1668
+ "learning_rate": 6.834285714285715e-06,
1669
+ "loss": 1.8765,
1670
+ "step": 277
1671
+ },
1672
+ {
1673
+ "epoch": 1.59,
1674
+ "learning_rate": 6.822857142857143e-06,
1675
+ "loss": 1.8971,
1676
+ "step": 278
1677
+ },
1678
+ {
1679
+ "epoch": 1.59,
1680
+ "learning_rate": 6.8114285714285725e-06,
1681
+ "loss": 1.9175,
1682
+ "step": 279
1683
+ },
1684
+ {
1685
+ "epoch": 1.6,
1686
+ "learning_rate": 6.800000000000001e-06,
1687
+ "loss": 1.8901,
1688
+ "step": 280
1689
+ },
1690
+ {
1691
+ "epoch": 1.6,
1692
+ "learning_rate": 6.7885714285714286e-06,
1693
+ "loss": 1.912,
1694
+ "step": 281
1695
+ },
1696
+ {
1697
+ "epoch": 1.61,
1698
+ "learning_rate": 6.777142857142858e-06,
1699
+ "loss": 1.8572,
1700
+ "step": 282
1701
+ },
1702
+ {
1703
+ "epoch": 1.61,
1704
+ "learning_rate": 6.765714285714286e-06,
1705
+ "loss": 1.8419,
1706
+ "step": 283
1707
+ },
1708
+ {
1709
+ "epoch": 1.62,
1710
+ "learning_rate": 6.754285714285715e-06,
1711
+ "loss": 1.8567,
1712
+ "step": 284
1713
+ },
1714
+ {
1715
+ "epoch": 1.63,
1716
+ "learning_rate": 6.742857142857143e-06,
1717
+ "loss": 1.9234,
1718
+ "step": 285
1719
+ },
1720
+ {
1721
+ "epoch": 1.63,
1722
+ "learning_rate": 6.7314285714285724e-06,
1723
+ "loss": 1.8159,
1724
+ "step": 286
1725
+ },
1726
+ {
1727
+ "epoch": 1.64,
1728
+ "learning_rate": 6.720000000000001e-06,
1729
+ "loss": 1.8715,
1730
+ "step": 287
1731
+ },
1732
+ {
1733
+ "epoch": 1.64,
1734
+ "learning_rate": 6.7085714285714285e-06,
1735
+ "loss": 1.8771,
1736
+ "step": 288
1737
+ },
1738
+ {
1739
+ "epoch": 1.65,
1740
+ "learning_rate": 6.697142857142858e-06,
1741
+ "loss": 1.836,
1742
+ "step": 289
1743
+ },
1744
+ {
1745
+ "epoch": 1.65,
1746
+ "learning_rate": 6.685714285714286e-06,
1747
+ "loss": 1.8798,
1748
+ "step": 290
1749
+ },
1750
+ {
1751
+ "epoch": 1.66,
1752
+ "learning_rate": 6.6742857142857155e-06,
1753
+ "loss": 1.9188,
1754
+ "step": 291
1755
+ },
1756
+ {
1757
+ "epoch": 1.67,
1758
+ "learning_rate": 6.662857142857143e-06,
1759
+ "loss": 1.8665,
1760
+ "step": 292
1761
+ },
1762
+ {
1763
+ "epoch": 1.67,
1764
+ "learning_rate": 6.651428571428572e-06,
1765
+ "loss": 1.8801,
1766
+ "step": 293
1767
+ },
1768
+ {
1769
+ "epoch": 1.68,
1770
+ "learning_rate": 6.640000000000001e-06,
1771
+ "loss": 1.8261,
1772
+ "step": 294
1773
+ },
1774
+ {
1775
+ "epoch": 1.68,
1776
+ "learning_rate": 6.628571428571428e-06,
1777
+ "loss": 1.8659,
1778
+ "step": 295
1779
+ },
1780
+ {
1781
+ "epoch": 1.69,
1782
+ "learning_rate": 6.617142857142858e-06,
1783
+ "loss": 1.9501,
1784
+ "step": 296
1785
+ },
1786
+ {
1787
+ "epoch": 1.69,
1788
+ "learning_rate": 6.605714285714286e-06,
1789
+ "loss": 1.8535,
1790
+ "step": 297
1791
+ },
1792
+ {
1793
+ "epoch": 1.7,
1794
+ "learning_rate": 6.594285714285715e-06,
1795
+ "loss": 1.8963,
1796
+ "step": 298
1797
+ },
1798
+ {
1799
+ "epoch": 1.71,
1800
+ "learning_rate": 6.582857142857143e-06,
1801
+ "loss": 1.8672,
1802
+ "step": 299
1803
+ },
1804
+ {
1805
+ "epoch": 1.71,
1806
+ "learning_rate": 6.571428571428572e-06,
1807
+ "loss": 1.8863,
1808
+ "step": 300
1809
+ },
1810
+ {
1811
+ "epoch": 1.72,
1812
+ "learning_rate": 6.560000000000001e-06,
1813
+ "loss": 1.8532,
1814
+ "step": 301
1815
+ },
1816
+ {
1817
+ "epoch": 1.72,
1818
+ "learning_rate": 6.548571428571428e-06,
1819
+ "loss": 1.8971,
1820
+ "step": 302
1821
+ },
1822
+ {
1823
+ "epoch": 1.73,
1824
+ "learning_rate": 6.537142857142858e-06,
1825
+ "loss": 1.8611,
1826
+ "step": 303
1827
+ },
1828
+ {
1829
+ "epoch": 1.73,
1830
+ "learning_rate": 6.525714285714286e-06,
1831
+ "loss": 1.825,
1832
+ "step": 304
1833
+ },
1834
+ {
1835
+ "epoch": 1.74,
1836
+ "learning_rate": 6.514285714285715e-06,
1837
+ "loss": 1.8707,
1838
+ "step": 305
1839
+ },
1840
+ {
1841
+ "epoch": 1.75,
1842
+ "learning_rate": 6.502857142857143e-06,
1843
+ "loss": 1.9209,
1844
+ "step": 306
1845
+ },
1846
+ {
1847
+ "epoch": 1.75,
1848
+ "learning_rate": 6.491428571428572e-06,
1849
+ "loss": 1.8413,
1850
+ "step": 307
1851
+ },
1852
+ {
1853
+ "epoch": 1.76,
1854
+ "learning_rate": 6.480000000000001e-06,
1855
+ "loss": 1.8566,
1856
+ "step": 308
1857
+ },
1858
+ {
1859
+ "epoch": 1.76,
1860
+ "learning_rate": 6.46857142857143e-06,
1861
+ "loss": 1.8916,
1862
+ "step": 309
1863
+ },
1864
+ {
1865
+ "epoch": 1.77,
1866
+ "learning_rate": 6.4571428571428575e-06,
1867
+ "loss": 1.9048,
1868
+ "step": 310
1869
+ },
1870
+ {
1871
+ "epoch": 1.77,
1872
+ "learning_rate": 6.445714285714286e-06,
1873
+ "loss": 1.8584,
1874
+ "step": 311
1875
+ },
1876
+ {
1877
+ "epoch": 1.78,
1878
+ "learning_rate": 6.434285714285715e-06,
1879
+ "loss": 1.8349,
1880
+ "step": 312
1881
+ },
1882
+ {
1883
+ "epoch": 1.79,
1884
+ "learning_rate": 6.422857142857143e-06,
1885
+ "loss": 1.89,
1886
+ "step": 313
1887
+ },
1888
+ {
1889
+ "epoch": 1.79,
1890
+ "learning_rate": 6.411428571428572e-06,
1891
+ "loss": 1.8845,
1892
+ "step": 314
1893
+ },
1894
+ {
1895
+ "epoch": 1.8,
1896
+ "learning_rate": 6.4000000000000006e-06,
1897
+ "loss": 1.7968,
1898
+ "step": 315
1899
+ },
1900
+ {
1901
+ "epoch": 1.8,
1902
+ "learning_rate": 6.38857142857143e-06,
1903
+ "loss": 1.8606,
1904
+ "step": 316
1905
+ },
1906
+ {
1907
+ "epoch": 1.81,
1908
+ "learning_rate": 6.3771428571428574e-06,
1909
+ "loss": 1.8538,
1910
+ "step": 317
1911
+ },
1912
+ {
1913
+ "epoch": 1.81,
1914
+ "learning_rate": 6.365714285714286e-06,
1915
+ "loss": 1.8271,
1916
+ "step": 318
1917
+ },
1918
+ {
1919
+ "epoch": 1.82,
1920
+ "learning_rate": 6.354285714285715e-06,
1921
+ "loss": 1.8151,
1922
+ "step": 319
1923
+ },
1924
+ {
1925
+ "epoch": 1.83,
1926
+ "learning_rate": 6.342857142857143e-06,
1927
+ "loss": 1.9135,
1928
+ "step": 320
1929
+ },
1930
+ {
1931
+ "epoch": 1.83,
1932
+ "learning_rate": 6.331428571428572e-06,
1933
+ "loss": 1.8722,
1934
+ "step": 321
1935
+ },
1936
+ {
1937
+ "epoch": 1.84,
1938
+ "learning_rate": 6.3200000000000005e-06,
1939
+ "loss": 1.9211,
1940
+ "step": 322
1941
+ },
1942
+ {
1943
+ "epoch": 1.84,
1944
+ "learning_rate": 6.30857142857143e-06,
1945
+ "loss": 1.9172,
1946
+ "step": 323
1947
+ },
1948
+ {
1949
+ "epoch": 1.85,
1950
+ "learning_rate": 6.297142857142857e-06,
1951
+ "loss": 1.9246,
1952
+ "step": 324
1953
+ },
1954
+ {
1955
+ "epoch": 1.85,
1956
+ "learning_rate": 6.285714285714286e-06,
1957
+ "loss": 1.8079,
1958
+ "step": 325
1959
+ },
1960
+ {
1961
+ "epoch": 1.86,
1962
+ "learning_rate": 6.274285714285715e-06,
1963
+ "loss": 1.9163,
1964
+ "step": 326
1965
+ },
1966
+ {
1967
+ "epoch": 1.87,
1968
+ "learning_rate": 6.2628571428571435e-06,
1969
+ "loss": 1.924,
1970
+ "step": 327
1971
+ },
1972
+ {
1973
+ "epoch": 1.87,
1974
+ "learning_rate": 6.251428571428572e-06,
1975
+ "loss": 1.8322,
1976
+ "step": 328
1977
+ },
1978
+ {
1979
+ "epoch": 1.88,
1980
+ "learning_rate": 6.24e-06,
1981
+ "loss": 1.9303,
1982
+ "step": 329
1983
+ },
1984
+ {
1985
+ "epoch": 1.88,
1986
+ "learning_rate": 6.22857142857143e-06,
1987
+ "loss": 1.9153,
1988
+ "step": 330
1989
+ },
1990
+ {
1991
+ "epoch": 1.89,
1992
+ "learning_rate": 6.217142857142857e-06,
1993
+ "loss": 1.9472,
1994
+ "step": 331
1995
+ },
1996
+ {
1997
+ "epoch": 1.89,
1998
+ "learning_rate": 6.205714285714286e-06,
1999
+ "loss": 1.8898,
2000
+ "step": 332
2001
+ },
2002
+ {
2003
+ "epoch": 1.9,
2004
+ "learning_rate": 6.194285714285715e-06,
2005
+ "loss": 1.8602,
2006
+ "step": 333
2007
+ },
2008
+ {
2009
+ "epoch": 1.91,
2010
+ "learning_rate": 6.1828571428571434e-06,
2011
+ "loss": 1.8837,
2012
+ "step": 334
2013
+ },
2014
+ {
2015
+ "epoch": 1.91,
2016
+ "learning_rate": 6.171428571428572e-06,
2017
+ "loss": 1.869,
2018
+ "step": 335
2019
+ },
2020
+ {
2021
+ "epoch": 1.92,
2022
+ "learning_rate": 6.16e-06,
2023
+ "loss": 1.8635,
2024
+ "step": 336
2025
+ },
2026
+ {
2027
+ "epoch": 1.92,
2028
+ "learning_rate": 6.14857142857143e-06,
2029
+ "loss": 1.8479,
2030
+ "step": 337
2031
+ },
2032
+ {
2033
+ "epoch": 1.93,
2034
+ "learning_rate": 6.137142857142858e-06,
2035
+ "loss": 1.842,
2036
+ "step": 338
2037
+ },
2038
+ {
2039
+ "epoch": 1.93,
2040
+ "learning_rate": 6.125714285714286e-06,
2041
+ "loss": 1.8694,
2042
+ "step": 339
2043
+ },
2044
+ {
2045
+ "epoch": 1.94,
2046
+ "learning_rate": 6.114285714285715e-06,
2047
+ "loss": 1.9413,
2048
+ "step": 340
2049
+ },
2050
+ {
2051
+ "epoch": 1.95,
2052
+ "learning_rate": 6.102857142857143e-06,
2053
+ "loss": 1.8717,
2054
+ "step": 341
2055
+ },
2056
+ {
2057
+ "epoch": 1.95,
2058
+ "learning_rate": 6.091428571428572e-06,
2059
+ "loss": 1.9343,
2060
+ "step": 342
2061
+ },
2062
+ {
2063
+ "epoch": 1.96,
2064
+ "learning_rate": 6.08e-06,
2065
+ "loss": 1.8665,
2066
+ "step": 343
2067
+ },
2068
+ {
2069
+ "epoch": 1.96,
2070
+ "learning_rate": 6.0685714285714295e-06,
2071
+ "loss": 1.8946,
2072
+ "step": 344
2073
+ },
2074
+ {
2075
+ "epoch": 1.97,
2076
+ "learning_rate": 6.057142857142858e-06,
2077
+ "loss": 1.8547,
2078
+ "step": 345
2079
+ },
2080
+ {
2081
+ "epoch": 1.97,
2082
+ "learning_rate": 6.0457142857142855e-06,
2083
+ "loss": 1.8472,
2084
+ "step": 346
2085
+ },
2086
+ {
2087
+ "epoch": 1.98,
2088
+ "learning_rate": 6.034285714285715e-06,
2089
+ "loss": 1.8862,
2090
+ "step": 347
2091
+ },
2092
+ {
2093
+ "epoch": 1.99,
2094
+ "learning_rate": 6.022857142857143e-06,
2095
+ "loss": 1.894,
2096
+ "step": 348
2097
+ },
2098
+ {
2099
+ "epoch": 1.99,
2100
+ "learning_rate": 6.011428571428572e-06,
2101
+ "loss": 1.8344,
2102
+ "step": 349
2103
+ },
2104
+ {
2105
+ "epoch": 2.0,
2106
+ "learning_rate": 6e-06,
2107
+ "loss": 1.8955,
2108
+ "step": 350
2109
+ },
2110
+ {
2111
+ "epoch": 2.0,
2112
+ "learning_rate": 5.9885714285714294e-06,
2113
+ "loss": 1.4813,
2114
+ "step": 351
2115
+ },
2116
+ {
2117
+ "epoch": 2.01,
2118
+ "learning_rate": 5.977142857142858e-06,
2119
+ "loss": 1.5577,
2120
+ "step": 352
2121
+ },
2122
+ {
2123
+ "epoch": 2.01,
2124
+ "learning_rate": 5.9657142857142855e-06,
2125
+ "loss": 1.5384,
2126
+ "step": 353
2127
+ },
2128
+ {
2129
+ "epoch": 2.02,
2130
+ "learning_rate": 5.954285714285715e-06,
2131
+ "loss": 1.5282,
2132
+ "step": 354
2133
+ },
2134
+ {
2135
+ "epoch": 2.03,
2136
+ "learning_rate": 5.942857142857143e-06,
2137
+ "loss": 1.607,
2138
+ "step": 355
2139
+ },
2140
+ {
2141
+ "epoch": 2.03,
2142
+ "learning_rate": 5.9314285714285725e-06,
2143
+ "loss": 1.5854,
2144
+ "step": 356
2145
+ },
2146
+ {
2147
+ "epoch": 2.04,
2148
+ "learning_rate": 5.92e-06,
2149
+ "loss": 1.5224,
2150
+ "step": 357
2151
+ },
2152
+ {
2153
+ "epoch": 2.04,
2154
+ "learning_rate": 5.908571428571429e-06,
2155
+ "loss": 1.5569,
2156
+ "step": 358
2157
+ },
2158
+ {
2159
+ "epoch": 2.05,
2160
+ "learning_rate": 5.897142857142858e-06,
2161
+ "loss": 1.517,
2162
+ "step": 359
2163
+ },
2164
+ {
2165
+ "epoch": 2.05,
2166
+ "learning_rate": 5.885714285714285e-06,
2167
+ "loss": 1.5195,
2168
+ "step": 360
2169
+ },
2170
+ {
2171
+ "epoch": 2.06,
2172
+ "learning_rate": 5.874285714285715e-06,
2173
+ "loss": 1.506,
2174
+ "step": 361
2175
+ },
2176
+ {
2177
+ "epoch": 2.07,
2178
+ "learning_rate": 5.862857142857143e-06,
2179
+ "loss": 1.5612,
2180
+ "step": 362
2181
+ },
2182
+ {
2183
+ "epoch": 2.07,
2184
+ "learning_rate": 5.851428571428572e-06,
2185
+ "loss": 1.5164,
2186
+ "step": 363
2187
+ },
2188
+ {
2189
+ "epoch": 2.08,
2190
+ "learning_rate": 5.84e-06,
2191
+ "loss": 1.5524,
2192
+ "step": 364
2193
+ },
2194
+ {
2195
+ "epoch": 2.08,
2196
+ "learning_rate": 5.828571428571429e-06,
2197
+ "loss": 1.4835,
2198
+ "step": 365
2199
+ },
2200
+ {
2201
+ "epoch": 2.09,
2202
+ "learning_rate": 5.817142857142858e-06,
2203
+ "loss": 1.4747,
2204
+ "step": 366
2205
+ },
2206
+ {
2207
+ "epoch": 2.09,
2208
+ "learning_rate": 5.805714285714285e-06,
2209
+ "loss": 1.5536,
2210
+ "step": 367
2211
+ },
2212
+ {
2213
+ "epoch": 2.1,
2214
+ "learning_rate": 5.794285714285715e-06,
2215
+ "loss": 1.5723,
2216
+ "step": 368
2217
+ },
2218
+ {
2219
+ "epoch": 2.11,
2220
+ "learning_rate": 5.782857142857143e-06,
2221
+ "loss": 1.5643,
2222
+ "step": 369
2223
+ },
2224
+ {
2225
+ "epoch": 2.11,
2226
+ "learning_rate": 5.771428571428572e-06,
2227
+ "loss": 1.5,
2228
+ "step": 370
2229
+ },
2230
+ {
2231
+ "epoch": 2.12,
2232
+ "learning_rate": 5.76e-06,
2233
+ "loss": 1.4934,
2234
+ "step": 371
2235
+ },
2236
+ {
2237
+ "epoch": 2.12,
2238
+ "learning_rate": 5.748571428571429e-06,
2239
+ "loss": 1.537,
2240
+ "step": 372
2241
+ },
2242
+ {
2243
+ "epoch": 2.13,
2244
+ "learning_rate": 5.737142857142858e-06,
2245
+ "loss": 1.5511,
2246
+ "step": 373
2247
+ },
2248
+ {
2249
+ "epoch": 2.13,
2250
+ "learning_rate": 5.725714285714287e-06,
2251
+ "loss": 1.5645,
2252
+ "step": 374
2253
+ },
2254
+ {
2255
+ "epoch": 2.14,
2256
+ "learning_rate": 5.7142857142857145e-06,
2257
+ "loss": 1.5472,
2258
+ "step": 375
2259
+ },
2260
+ {
2261
+ "epoch": 2.15,
2262
+ "learning_rate": 5.702857142857143e-06,
2263
+ "loss": 1.5441,
2264
+ "step": 376
2265
+ },
2266
+ {
2267
+ "epoch": 2.15,
2268
+ "learning_rate": 5.691428571428572e-06,
2269
+ "loss": 1.5462,
2270
+ "step": 377
2271
+ },
2272
+ {
2273
+ "epoch": 2.16,
2274
+ "learning_rate": 5.68e-06,
2275
+ "loss": 1.5596,
2276
+ "step": 378
2277
+ },
2278
+ {
2279
+ "epoch": 2.16,
2280
+ "learning_rate": 5.668571428571429e-06,
2281
+ "loss": 1.4952,
2282
+ "step": 379
2283
+ },
2284
+ {
2285
+ "epoch": 2.17,
2286
+ "learning_rate": 5.6571428571428576e-06,
2287
+ "loss": 1.5553,
2288
+ "step": 380
2289
+ },
2290
+ {
2291
+ "epoch": 2.17,
2292
+ "learning_rate": 5.645714285714287e-06,
2293
+ "loss": 1.6068,
2294
+ "step": 381
2295
+ },
2296
+ {
2297
+ "epoch": 2.18,
2298
+ "learning_rate": 5.6342857142857144e-06,
2299
+ "loss": 1.5156,
2300
+ "step": 382
2301
+ },
2302
+ {
2303
+ "epoch": 2.19,
2304
+ "learning_rate": 5.622857142857143e-06,
2305
+ "loss": 1.5715,
2306
+ "step": 383
2307
+ },
2308
+ {
2309
+ "epoch": 2.19,
2310
+ "learning_rate": 5.611428571428572e-06,
2311
+ "loss": 1.5572,
2312
+ "step": 384
2313
+ },
2314
+ {
2315
+ "epoch": 2.2,
2316
+ "learning_rate": 5.600000000000001e-06,
2317
+ "loss": 1.4732,
2318
+ "step": 385
2319
+ },
2320
+ {
2321
+ "epoch": 2.2,
2322
+ "learning_rate": 5.588571428571429e-06,
2323
+ "loss": 1.4676,
2324
+ "step": 386
2325
+ },
2326
+ {
2327
+ "epoch": 2.21,
2328
+ "learning_rate": 5.5771428571428575e-06,
2329
+ "loss": 1.4884,
2330
+ "step": 387
2331
+ },
2332
+ {
2333
+ "epoch": 2.21,
2334
+ "learning_rate": 5.565714285714287e-06,
2335
+ "loss": 1.5728,
2336
+ "step": 388
2337
+ },
2338
+ {
2339
+ "epoch": 2.22,
2340
+ "learning_rate": 5.554285714285714e-06,
2341
+ "loss": 1.4687,
2342
+ "step": 389
2343
+ },
2344
+ {
2345
+ "epoch": 2.23,
2346
+ "learning_rate": 5.542857142857143e-06,
2347
+ "loss": 1.5328,
2348
+ "step": 390
2349
+ },
2350
+ {
2351
+ "epoch": 2.23,
2352
+ "learning_rate": 5.531428571428572e-06,
2353
+ "loss": 1.5801,
2354
+ "step": 391
2355
+ },
2356
+ {
2357
+ "epoch": 2.24,
2358
+ "learning_rate": 5.5200000000000005e-06,
2359
+ "loss": 1.5507,
2360
+ "step": 392
2361
+ },
2362
+ {
2363
+ "epoch": 2.24,
2364
+ "learning_rate": 5.508571428571429e-06,
2365
+ "loss": 1.4823,
2366
+ "step": 393
2367
+ },
2368
+ {
2369
+ "epoch": 2.25,
2370
+ "learning_rate": 5.497142857142857e-06,
2371
+ "loss": 1.5014,
2372
+ "step": 394
2373
+ },
2374
+ {
2375
+ "epoch": 2.25,
2376
+ "learning_rate": 5.485714285714287e-06,
2377
+ "loss": 1.5224,
2378
+ "step": 395
2379
+ },
2380
+ {
2381
+ "epoch": 2.26,
2382
+ "learning_rate": 5.474285714285714e-06,
2383
+ "loss": 1.6566,
2384
+ "step": 396
2385
+ },
2386
+ {
2387
+ "epoch": 2.27,
2388
+ "learning_rate": 5.462857142857143e-06,
2389
+ "loss": 1.5042,
2390
+ "step": 397
2391
+ },
2392
+ {
2393
+ "epoch": 2.27,
2394
+ "learning_rate": 5.451428571428572e-06,
2395
+ "loss": 1.5212,
2396
+ "step": 398
2397
+ },
2398
+ {
2399
+ "epoch": 2.28,
2400
+ "learning_rate": 5.4400000000000004e-06,
2401
+ "loss": 1.4854,
2402
+ "step": 399
2403
+ },
2404
+ {
2405
+ "epoch": 2.28,
2406
+ "learning_rate": 5.428571428571429e-06,
2407
+ "loss": 1.5316,
2408
+ "step": 400
2409
+ },
2410
+ {
2411
+ "epoch": 2.29,
2412
+ "learning_rate": 5.417142857142857e-06,
2413
+ "loss": 1.4371,
2414
+ "step": 401
2415
+ },
2416
+ {
2417
+ "epoch": 2.29,
2418
+ "learning_rate": 5.405714285714287e-06,
2419
+ "loss": 1.5148,
2420
+ "step": 402
2421
+ },
2422
+ {
2423
+ "epoch": 2.3,
2424
+ "learning_rate": 5.394285714285715e-06,
2425
+ "loss": 1.479,
2426
+ "step": 403
2427
+ },
2428
+ {
2429
+ "epoch": 2.31,
2430
+ "learning_rate": 5.382857142857143e-06,
2431
+ "loss": 1.5122,
2432
+ "step": 404
2433
+ },
2434
+ {
2435
+ "epoch": 2.31,
2436
+ "learning_rate": 5.371428571428572e-06,
2437
+ "loss": 1.5215,
2438
+ "step": 405
2439
+ },
2440
+ {
2441
+ "epoch": 2.32,
2442
+ "learning_rate": 5.36e-06,
2443
+ "loss": 1.5144,
2444
+ "step": 406
2445
+ },
2446
+ {
2447
+ "epoch": 2.32,
2448
+ "learning_rate": 5.348571428571429e-06,
2449
+ "loss": 1.7639,
2450
+ "step": 407
2451
+ },
2452
+ {
2453
+ "epoch": 2.33,
2454
+ "learning_rate": 5.337142857142857e-06,
2455
+ "loss": 1.4448,
2456
+ "step": 408
2457
+ },
2458
+ {
2459
+ "epoch": 2.33,
2460
+ "learning_rate": 5.3257142857142865e-06,
2461
+ "loss": 1.7201,
2462
+ "step": 409
2463
+ },
2464
+ {
2465
+ "epoch": 2.34,
2466
+ "learning_rate": 5.314285714285715e-06,
2467
+ "loss": 1.4932,
2468
+ "step": 410
2469
+ },
2470
+ {
2471
+ "epoch": 2.35,
2472
+ "learning_rate": 5.3028571428571425e-06,
2473
+ "loss": 1.4845,
2474
+ "step": 411
2475
+ },
2476
+ {
2477
+ "epoch": 2.35,
2478
+ "learning_rate": 5.291428571428572e-06,
2479
+ "loss": 1.4798,
2480
+ "step": 412
2481
+ },
2482
+ {
2483
+ "epoch": 2.36,
2484
+ "learning_rate": 5.28e-06,
2485
+ "loss": 1.4747,
2486
+ "step": 413
2487
+ },
2488
+ {
2489
+ "epoch": 2.36,
2490
+ "learning_rate": 5.268571428571429e-06,
2491
+ "loss": 1.4672,
2492
+ "step": 414
2493
+ },
2494
+ {
2495
+ "epoch": 2.37,
2496
+ "learning_rate": 5.257142857142857e-06,
2497
+ "loss": 1.4649,
2498
+ "step": 415
2499
+ },
2500
+ {
2501
+ "epoch": 2.37,
2502
+ "learning_rate": 5.2457142857142864e-06,
2503
+ "loss": 1.4545,
2504
+ "step": 416
2505
+ },
2506
+ {
2507
+ "epoch": 2.38,
2508
+ "learning_rate": 5.234285714285715e-06,
2509
+ "loss": 1.4801,
2510
+ "step": 417
2511
+ },
2512
+ {
2513
+ "epoch": 2.39,
2514
+ "learning_rate": 5.2228571428571425e-06,
2515
+ "loss": 1.5474,
2516
+ "step": 418
2517
+ },
2518
+ {
2519
+ "epoch": 2.39,
2520
+ "learning_rate": 5.211428571428572e-06,
2521
+ "loss": 1.4755,
2522
+ "step": 419
2523
+ },
2524
+ {
2525
+ "epoch": 2.4,
2526
+ "learning_rate": 5.2e-06,
2527
+ "loss": 1.5061,
2528
+ "step": 420
2529
+ },
2530
+ {
2531
+ "epoch": 2.4,
2532
+ "learning_rate": 5.1885714285714295e-06,
2533
+ "loss": 1.4828,
2534
+ "step": 421
2535
+ },
2536
+ {
2537
+ "epoch": 2.41,
2538
+ "learning_rate": 5.177142857142857e-06,
2539
+ "loss": 1.4669,
2540
+ "step": 422
2541
+ },
2542
+ {
2543
+ "epoch": 2.41,
2544
+ "learning_rate": 5.165714285714286e-06,
2545
+ "loss": 1.4596,
2546
+ "step": 423
2547
+ },
2548
+ {
2549
+ "epoch": 2.42,
2550
+ "learning_rate": 5.154285714285715e-06,
2551
+ "loss": 1.5299,
2552
+ "step": 424
2553
+ },
2554
+ {
2555
+ "epoch": 2.43,
2556
+ "learning_rate": 5.142857142857142e-06,
2557
+ "loss": 1.4452,
2558
+ "step": 425
2559
+ },
2560
+ {
2561
+ "epoch": 2.43,
2562
+ "learning_rate": 5.131428571428572e-06,
2563
+ "loss": 1.5066,
2564
+ "step": 426
2565
+ },
2566
+ {
2567
+ "epoch": 2.44,
2568
+ "learning_rate": 5.12e-06,
2569
+ "loss": 1.5446,
2570
+ "step": 427
2571
+ },
2572
+ {
2573
+ "epoch": 2.44,
2574
+ "learning_rate": 5.108571428571429e-06,
2575
+ "loss": 1.4556,
2576
+ "step": 428
2577
+ },
2578
+ {
2579
+ "epoch": 2.45,
2580
+ "learning_rate": 5.097142857142857e-06,
2581
+ "loss": 1.4567,
2582
+ "step": 429
2583
+ },
2584
+ {
2585
+ "epoch": 2.45,
2586
+ "learning_rate": 5.085714285714286e-06,
2587
+ "loss": 1.5352,
2588
+ "step": 430
2589
+ },
2590
+ {
2591
+ "epoch": 2.46,
2592
+ "learning_rate": 5.074285714285715e-06,
2593
+ "loss": 1.5015,
2594
+ "step": 431
2595
+ },
2596
+ {
2597
+ "epoch": 2.47,
2598
+ "learning_rate": 5.062857142857144e-06,
2599
+ "loss": 1.4837,
2600
+ "step": 432
2601
+ },
2602
+ {
2603
+ "epoch": 2.47,
2604
+ "learning_rate": 5.051428571428572e-06,
2605
+ "loss": 1.5209,
2606
+ "step": 433
2607
+ },
2608
+ {
2609
+ "epoch": 2.48,
2610
+ "learning_rate": 5.04e-06,
2611
+ "loss": 1.4885,
2612
+ "step": 434
2613
+ },
2614
+ {
2615
+ "epoch": 2.48,
2616
+ "learning_rate": 5.028571428571429e-06,
2617
+ "loss": 1.6766,
2618
+ "step": 435
2619
+ },
2620
+ {
2621
+ "epoch": 2.49,
2622
+ "learning_rate": 5.017142857142857e-06,
2623
+ "loss": 1.6379,
2624
+ "step": 436
2625
+ },
2626
+ {
2627
+ "epoch": 2.49,
2628
+ "learning_rate": 5.005714285714286e-06,
2629
+ "loss": 1.4389,
2630
+ "step": 437
2631
+ },
2632
+ {
2633
+ "epoch": 2.5,
2634
+ "learning_rate": 4.994285714285715e-06,
2635
+ "loss": 1.4618,
2636
+ "step": 438
2637
+ },
2638
+ {
2639
+ "epoch": 2.5,
2640
+ "learning_rate": 4.982857142857143e-06,
2641
+ "loss": 1.5679,
2642
+ "step": 439
2643
+ },
2644
+ {
2645
+ "epoch": 2.51,
2646
+ "learning_rate": 4.971428571428572e-06,
2647
+ "loss": 1.5112,
2648
+ "step": 440
2649
+ },
2650
+ {
2651
+ "epoch": 2.52,
2652
+ "learning_rate": 4.960000000000001e-06,
2653
+ "loss": 1.5533,
2654
+ "step": 441
2655
+ },
2656
+ {
2657
+ "epoch": 2.52,
2658
+ "learning_rate": 4.948571428571429e-06,
2659
+ "loss": 1.4795,
2660
+ "step": 442
2661
+ },
2662
+ {
2663
+ "epoch": 2.53,
2664
+ "learning_rate": 4.937142857142858e-06,
2665
+ "loss": 1.4529,
2666
+ "step": 443
2667
+ },
2668
+ {
2669
+ "epoch": 2.53,
2670
+ "learning_rate": 4.925714285714286e-06,
2671
+ "loss": 1.4912,
2672
+ "step": 444
2673
+ },
2674
+ {
2675
+ "epoch": 2.54,
2676
+ "learning_rate": 4.9142857142857145e-06,
2677
+ "loss": 1.4789,
2678
+ "step": 445
2679
+ },
2680
+ {
2681
+ "epoch": 2.54,
2682
+ "learning_rate": 4.902857142857143e-06,
2683
+ "loss": 1.4331,
2684
+ "step": 446
2685
+ },
2686
+ {
2687
+ "epoch": 2.55,
2688
+ "learning_rate": 4.891428571428572e-06,
2689
+ "loss": 1.4892,
2690
+ "step": 447
2691
+ },
2692
+ {
2693
+ "epoch": 2.56,
2694
+ "learning_rate": 4.880000000000001e-06,
2695
+ "loss": 1.4924,
2696
+ "step": 448
2697
+ },
2698
+ {
2699
+ "epoch": 2.56,
2700
+ "learning_rate": 4.868571428571429e-06,
2701
+ "loss": 1.459,
2702
+ "step": 449
2703
+ },
2704
+ {
2705
+ "epoch": 2.57,
2706
+ "learning_rate": 4.857142857142858e-06,
2707
+ "loss": 1.5136,
2708
+ "step": 450
2709
+ },
2710
+ {
2711
+ "epoch": 2.57,
2712
+ "learning_rate": 4.845714285714286e-06,
2713
+ "loss": 1.4582,
2714
+ "step": 451
2715
+ },
2716
+ {
2717
+ "epoch": 2.58,
2718
+ "learning_rate": 4.8342857142857145e-06,
2719
+ "loss": 1.4864,
2720
+ "step": 452
2721
+ },
2722
+ {
2723
+ "epoch": 2.58,
2724
+ "learning_rate": 4.822857142857143e-06,
2725
+ "loss": 1.4838,
2726
+ "step": 453
2727
+ },
2728
+ {
2729
+ "epoch": 2.59,
2730
+ "learning_rate": 4.811428571428572e-06,
2731
+ "loss": 1.4838,
2732
+ "step": 454
2733
+ },
2734
+ {
2735
+ "epoch": 2.6,
2736
+ "learning_rate": 4.800000000000001e-06,
2737
+ "loss": 1.5041,
2738
+ "step": 455
2739
+ },
2740
+ {
2741
+ "epoch": 2.6,
2742
+ "learning_rate": 4.788571428571429e-06,
2743
+ "loss": 1.4823,
2744
+ "step": 456
2745
+ },
2746
+ {
2747
+ "epoch": 2.61,
2748
+ "learning_rate": 4.7771428571428575e-06,
2749
+ "loss": 1.492,
2750
+ "step": 457
2751
+ },
2752
+ {
2753
+ "epoch": 2.61,
2754
+ "learning_rate": 4.765714285714286e-06,
2755
+ "loss": 1.4655,
2756
+ "step": 458
2757
+ },
2758
+ {
2759
+ "epoch": 2.62,
2760
+ "learning_rate": 4.754285714285714e-06,
2761
+ "loss": 1.4767,
2762
+ "step": 459
2763
+ },
2764
+ {
2765
+ "epoch": 2.62,
2766
+ "learning_rate": 4.742857142857144e-06,
2767
+ "loss": 1.462,
2768
+ "step": 460
2769
+ },
2770
+ {
2771
+ "epoch": 2.63,
2772
+ "learning_rate": 4.731428571428572e-06,
2773
+ "loss": 1.4796,
2774
+ "step": 461
2775
+ },
2776
+ {
2777
+ "epoch": 2.64,
2778
+ "learning_rate": 4.7200000000000005e-06,
2779
+ "loss": 1.5112,
2780
+ "step": 462
2781
+ },
2782
+ {
2783
+ "epoch": 2.64,
2784
+ "learning_rate": 4.708571428571429e-06,
2785
+ "loss": 1.4453,
2786
+ "step": 463
2787
+ },
2788
+ {
2789
+ "epoch": 2.65,
2790
+ "learning_rate": 4.6971428571428574e-06,
2791
+ "loss": 1.4278,
2792
+ "step": 464
2793
+ },
2794
+ {
2795
+ "epoch": 2.65,
2796
+ "learning_rate": 4.685714285714286e-06,
2797
+ "loss": 1.4257,
2798
+ "step": 465
2799
+ },
2800
+ {
2801
+ "epoch": 2.66,
2802
+ "learning_rate": 4.674285714285714e-06,
2803
+ "loss": 1.4736,
2804
+ "step": 466
2805
+ },
2806
+ {
2807
+ "epoch": 2.66,
2808
+ "learning_rate": 4.662857142857144e-06,
2809
+ "loss": 1.5285,
2810
+ "step": 467
2811
+ },
2812
+ {
2813
+ "epoch": 2.67,
2814
+ "learning_rate": 4.651428571428572e-06,
2815
+ "loss": 1.507,
2816
+ "step": 468
2817
+ },
2818
+ {
2819
+ "epoch": 2.68,
2820
+ "learning_rate": 4.6400000000000005e-06,
2821
+ "loss": 1.5101,
2822
+ "step": 469
2823
+ },
2824
+ {
2825
+ "epoch": 2.68,
2826
+ "learning_rate": 4.628571428571429e-06,
2827
+ "loss": 1.4296,
2828
+ "step": 470
2829
+ },
2830
+ {
2831
+ "epoch": 2.69,
2832
+ "learning_rate": 4.617142857142857e-06,
2833
+ "loss": 1.4725,
2834
+ "step": 471
2835
+ },
2836
+ {
2837
+ "epoch": 2.69,
2838
+ "learning_rate": 4.605714285714286e-06,
2839
+ "loss": 1.4013,
2840
+ "step": 472
2841
+ },
2842
+ {
2843
+ "epoch": 2.7,
2844
+ "learning_rate": 4.594285714285714e-06,
2845
+ "loss": 1.5,
2846
+ "step": 473
2847
+ },
2848
+ {
2849
+ "epoch": 2.7,
2850
+ "learning_rate": 4.5828571428571435e-06,
2851
+ "loss": 1.4106,
2852
+ "step": 474
2853
+ },
2854
+ {
2855
+ "epoch": 2.71,
2856
+ "learning_rate": 4.571428571428572e-06,
2857
+ "loss": 1.4365,
2858
+ "step": 475
2859
+ },
2860
+ {
2861
+ "epoch": 2.72,
2862
+ "learning_rate": 4.56e-06,
2863
+ "loss": 1.4955,
2864
+ "step": 476
2865
+ },
2866
+ {
2867
+ "epoch": 2.72,
2868
+ "learning_rate": 4.548571428571429e-06,
2869
+ "loss": 1.5075,
2870
+ "step": 477
2871
+ },
2872
+ {
2873
+ "epoch": 2.73,
2874
+ "learning_rate": 4.537142857142858e-06,
2875
+ "loss": 1.4233,
2876
+ "step": 478
2877
+ },
2878
+ {
2879
+ "epoch": 2.73,
2880
+ "learning_rate": 4.525714285714286e-06,
2881
+ "loss": 1.4557,
2882
+ "step": 479
2883
+ },
2884
+ {
2885
+ "epoch": 2.74,
2886
+ "learning_rate": 4.514285714285714e-06,
2887
+ "loss": 1.4704,
2888
+ "step": 480
2889
+ },
2890
+ {
2891
+ "epoch": 2.74,
2892
+ "learning_rate": 4.5028571428571434e-06,
2893
+ "loss": 1.4758,
2894
+ "step": 481
2895
+ },
2896
+ {
2897
+ "epoch": 2.75,
2898
+ "learning_rate": 4.491428571428572e-06,
2899
+ "loss": 1.4608,
2900
+ "step": 482
2901
+ },
2902
+ {
2903
+ "epoch": 2.76,
2904
+ "learning_rate": 4.48e-06,
2905
+ "loss": 1.5033,
2906
+ "step": 483
2907
+ },
2908
+ {
2909
+ "epoch": 2.76,
2910
+ "learning_rate": 4.468571428571429e-06,
2911
+ "loss": 1.4485,
2912
+ "step": 484
2913
+ },
2914
+ {
2915
+ "epoch": 2.77,
2916
+ "learning_rate": 4.457142857142858e-06,
2917
+ "loss": 1.4785,
2918
+ "step": 485
2919
+ },
2920
+ {
2921
+ "epoch": 2.77,
2922
+ "learning_rate": 4.445714285714286e-06,
2923
+ "loss": 1.4599,
2924
+ "step": 486
2925
+ },
2926
+ {
2927
+ "epoch": 2.78,
2928
+ "learning_rate": 4.434285714285715e-06,
2929
+ "loss": 1.443,
2930
+ "step": 487
2931
+ },
2932
+ {
2933
+ "epoch": 2.78,
2934
+ "learning_rate": 4.422857142857143e-06,
2935
+ "loss": 1.4236,
2936
+ "step": 488
2937
+ },
2938
+ {
2939
+ "epoch": 2.79,
2940
+ "learning_rate": 4.411428571428572e-06,
2941
+ "loss": 1.4812,
2942
+ "step": 489
2943
+ },
2944
+ {
2945
+ "epoch": 2.8,
2946
+ "learning_rate": 4.4e-06,
2947
+ "loss": 1.471,
2948
+ "step": 490
2949
+ },
2950
+ {
2951
+ "epoch": 2.8,
2952
+ "learning_rate": 4.388571428571429e-06,
2953
+ "loss": 1.4846,
2954
+ "step": 491
2955
+ },
2956
+ {
2957
+ "epoch": 2.81,
2958
+ "learning_rate": 4.377142857142858e-06,
2959
+ "loss": 1.7677,
2960
+ "step": 492
2961
+ },
2962
+ {
2963
+ "epoch": 2.81,
2964
+ "learning_rate": 4.3657142857142855e-06,
2965
+ "loss": 1.4363,
2966
+ "step": 493
2967
+ },
2968
+ {
2969
+ "epoch": 2.82,
2970
+ "learning_rate": 4.354285714285715e-06,
2971
+ "loss": 1.4643,
2972
+ "step": 494
2973
+ },
2974
+ {
2975
+ "epoch": 2.82,
2976
+ "learning_rate": 4.342857142857143e-06,
2977
+ "loss": 1.4879,
2978
+ "step": 495
2979
+ },
2980
+ {
2981
+ "epoch": 2.83,
2982
+ "learning_rate": 4.331428571428572e-06,
2983
+ "loss": 1.4658,
2984
+ "step": 496
2985
+ },
2986
+ {
2987
+ "epoch": 2.84,
2988
+ "learning_rate": 4.32e-06,
2989
+ "loss": 1.4439,
2990
+ "step": 497
2991
+ },
2992
+ {
2993
+ "epoch": 2.84,
2994
+ "learning_rate": 4.3085714285714294e-06,
2995
+ "loss": 1.4414,
2996
+ "step": 498
2997
+ },
2998
+ {
2999
+ "epoch": 2.85,
3000
+ "learning_rate": 4.297142857142858e-06,
3001
+ "loss": 1.454,
3002
+ "step": 499
3003
+ },
3004
+ {
3005
+ "epoch": 2.85,
3006
+ "learning_rate": 4.2857142857142855e-06,
3007
+ "loss": 1.473,
3008
+ "step": 500
3009
+ }
3010
+ ],
3011
+ "max_steps": 875,
3012
+ "num_train_epochs": 5,
3013
+ "total_flos": 2.869610866016256e+17,
3014
+ "trial_name": null,
3015
+ "trial_params": null
3016
+ }
checkpoint-500/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a0ca411a309a01320e37bb8a6b041d097c73a99f9df0b411d38617d2d3edb7c0
3
+ size 4143
config.json ADDED
@@ -0,0 +1,32 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "/notebooks/seq_2_seq-odqa/trial-1/mt5-large-2",
3
+ "architectures": [
4
+ "MT5ForConditionalGeneration"
5
+ ],
6
+ "d_ff": 2816,
7
+ "d_kv": 64,
8
+ "d_model": 1024,
9
+ "decoder_start_token_id": 0,
10
+ "dense_act_fn": "gelu_new",
11
+ "dropout_rate": 0.1,
12
+ "eos_token_id": 1,
13
+ "feed_forward_proj": "gated-gelu",
14
+ "initializer_factor": 1.0,
15
+ "is_encoder_decoder": true,
16
+ "is_gated_act": true,
17
+ "layer_norm_epsilon": 1e-06,
18
+ "model_type": "mt5",
19
+ "num_decoder_layers": 24,
20
+ "num_heads": 16,
21
+ "num_layers": 24,
22
+ "output_past": true,
23
+ "pad_token_id": 0,
24
+ "relative_attention_max_distance": 128,
25
+ "relative_attention_num_buckets": 32,
26
+ "tie_word_embeddings": false,
27
+ "tokenizer_class": "T5Tokenizer",
28
+ "torch_dtype": "float32",
29
+ "transformers_version": "4.31.0.dev0",
30
+ "use_cache": true,
31
+ "vocab_size": 250112
32
+ }
eval_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 4.99,
3
+ "eval_loss": 2.4137141704559326,
4
+ "eval_runtime": 177.0027,
5
+ "eval_samples": 19485,
6
+ "eval_samples_per_second": 110.083,
7
+ "eval_steps_per_second": 13.763
8
+ }
generation_config.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "decoder_start_token_id": 0,
4
+ "eos_token_id": 1,
5
+ "pad_token_id": 0,
6
+ "transformers_version": "4.31.0.dev0"
7
+ }
pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:29b96d40b893b892aca243a5727af9c8f00eb17a31a3e3c12fbdead81385439e
3
+ size 4918515385
runs/Jun14_15-16-51_nr49739vda/events.out.tfevents.1686755936.nr49739vda.311.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ed5dc7dc7ce3fef8f0aabacea707e8a933f31157a74cf55c664b8a420b0b5c5e
3
+ size 141756
runs/Jun14_15-16-51_nr49739vda/events.out.tfevents.1686762234.nr49739vda.311.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6dbd2f518eab82acc50c67a0ac2c8acda08ebd79d7e76107f714d65c4c6a7ec7
3
+ size 311
special_tokens_map.json ADDED
@@ -0,0 +1,5 @@
 
 
 
 
 
 
1
+ {
2
+ "eos_token": "</s>",
3
+ "pad_token": "<pad>",
4
+ "unk_token": "<unk>"
5
+ }
spiece.model ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ef78f86560d809067d12bac6c09f19a462cb3af3f54d2b8acbba26e1433125d6
3
+ size 4309802
tokenizer.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6a2367c5e6ea9fd74eb410302c1e4a85d8cc1a017d68ae472c7facdc60d95c32
3
+ size 16315384
tokenizer_config.json ADDED
@@ -0,0 +1,11 @@
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "additional_special_tokens": null,
3
+ "clean_up_tokenization_spaces": true,
4
+ "eos_token": "</s>",
5
+ "extra_ids": 0,
6
+ "model_max_length": 1000000000000000019884624838656,
7
+ "pad_token": "<pad>",
8
+ "sp_model_kwargs": {},
9
+ "tokenizer_class": "T5Tokenizer",
10
+ "unk_token": "<unk>"
11
+ }
train_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 4.99,
3
+ "train_loss": 1.6896061137063163,
4
+ "train_runtime": 6113.9003,
5
+ "train_samples": 44836,
6
+ "train_samples_per_second": 36.667,
7
+ "train_steps_per_second": 0.143
8
+ }
trainer_state.json ADDED
The diff for this file is too large to render. See raw diff
 
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a0ca411a309a01320e37bb8a6b041d097c73a99f9df0b411d38617d2d3edb7c0
3
+ size 4143