End of training
Browse files
.gitignore
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
checkpoint-*/
|
config.json
ADDED
@@ -0,0 +1,28 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "facebook/opt-350m",
|
3 |
+
"activation_dropout": 0.0,
|
4 |
+
"activation_function": "relu",
|
5 |
+
"architectures": [
|
6 |
+
"OPTForCausalLM"
|
7 |
+
],
|
8 |
+
"attention_dropout": 0.0,
|
9 |
+
"bos_token_id": 0,
|
10 |
+
"do_layer_norm_before": false,
|
11 |
+
"dropout": 0.1,
|
12 |
+
"eos_token_id": 2,
|
13 |
+
"ffn_dim": 4096,
|
14 |
+
"hidden_size": 1024,
|
15 |
+
"init_std": 0.02,
|
16 |
+
"layerdrop": 0.0,
|
17 |
+
"max_position_embeddings": 2048,
|
18 |
+
"model_type": "opt",
|
19 |
+
"num_attention_heads": 16,
|
20 |
+
"num_hidden_layers": 24,
|
21 |
+
"pad_token_id": 1,
|
22 |
+
"prefix": "</s>",
|
23 |
+
"torch_dtype": "float32",
|
24 |
+
"transformers_version": "4.19.2",
|
25 |
+
"use_cache": true,
|
26 |
+
"vocab_size": 50272,
|
27 |
+
"word_embed_proj_dim": 512
|
28 |
+
}
|
pytorch_model.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:7b7d98af091adf444d12af01f3d4f28b5b6fe6a51f680dd09100bcdf8133965c
|
3 |
+
size 1324908921
|
runs/May21_18-31-57_b3ffa722faa5/1653157942.9248893/events.out.tfevents.1653157942.b3ffa722faa5.59.1
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:b5eb70560fdb3c4fc15dd594751c6905a5ea3431c9e6472ab25b3e3084d50c3a
|
3 |
+
size 5193
|
runs/May21_18-31-57_b3ffa722faa5/events.out.tfevents.1653157942.b3ffa722faa5.59.0
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:39eeec56381a6002979e5cc185fbe1a64958efac7afaa99cdc91679622a91241
|
3 |
+
size 4359
|
training_args.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:d521556c1bed3b7f49d227325a73350e827d6e5e1db7d241492f15868b263982
|
3 |
+
size 3247
|