Commit
·
d732f01
1
Parent(s):
ce6601b
first commit
Browse files- config.json +26 -0
- merges.txt +0 -0
- optimizer.pt +3 -0
- pytorch_model.bin +3 -0
- scheduler.pt +3 -0
- special_tokens_map.json +1 -0
- tokenizer_config.json +1 -0
- trainer_state.json +1237 -0
- training_args.bin +3 -0
- vocab.json +0 -0
config.json
ADDED
|
@@ -0,0 +1,26 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "../declutr/EManuals_Roberta",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"LecbertForPreTraining"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.1,
|
| 7 |
+
"bos_token_id": 0,
|
| 8 |
+
"eos_token_id": 2,
|
| 9 |
+
"gradient_checkpointing": false,
|
| 10 |
+
"hidden_act": "gelu",
|
| 11 |
+
"hidden_dropout_prob": 0.1,
|
| 12 |
+
"hidden_size": 768,
|
| 13 |
+
"initializer_range": 0.02,
|
| 14 |
+
"intermediate_size": 3072,
|
| 15 |
+
"layer_norm_eps": 1e-05,
|
| 16 |
+
"max_position_embeddings": 514,
|
| 17 |
+
"model_type": "roberta",
|
| 18 |
+
"num_attention_heads": 12,
|
| 19 |
+
"num_hidden_layers": 12,
|
| 20 |
+
"pad_token_id": 1,
|
| 21 |
+
"position_embedding_type": "absolute",
|
| 22 |
+
"total_flos": 4462005349122048000,
|
| 23 |
+
"transformers_version": "4.7.0",
|
| 24 |
+
"type_vocab_size": 1,
|
| 25 |
+
"vocab_size": 50265
|
| 26 |
+
}
|
merges.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3022cd929280c149626dcfa39eb621b2cdd1f9c2e828a3f1bc4ba47d74334bb1
|
| 3 |
+
size 1007192453
|
pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:61a47fecd2be43c751eb350af0adf17af2d8dba335b16f13136916c2886e46fb
|
| 3 |
+
size 503624419
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a3391ffaeb57e510b8dab89a97fbce7976d101ea5597e6d9198b05b7c188c49b
|
| 3 |
+
size 559
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
{"bos_token": "<s>", "eos_token": "</s>", "unk_token": "<unk>", "sep_token": "</s>", "pad_token": "<pad>", "cls_token": "<s>", "mask_token": {"content": "<mask>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": false}}
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
{"unk_token": {"content": "<unk>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "__type": "AddedToken"}, "bos_token": {"content": "<s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "__type": "AddedToken"}, "eos_token": {"content": "</s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "__type": "AddedToken"}, "add_prefix_space": false, "errors": "replace", "sep_token": {"content": "</s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "__type": "AddedToken"}, "cls_token": {"content": "<s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "__type": "AddedToken"}, "pad_token": {"content": "<pad>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "__type": "AddedToken"}, "mask_token": {"content": "<mask>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "__type": "AddedToken"}, "model_max_length": 512, "special_tokens_map_file": "roberta-base-tok/special_tokens_map.json", "full_tokenizer_file": null, "name_or_path": "/scratch/j20200059/declutr/EManuals_Roberta", "add_special_tokens": false}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,1237 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.15092095071321246,
|
| 5 |
+
"global_step": 204000,
|
| 6 |
+
"is_local_process_zero": true,
|
| 7 |
+
"is_world_process_zero": true,
|
| 8 |
+
"log_history": [
|
| 9 |
+
{
|
| 10 |
+
"epoch": 0.000739808581927512,
|
| 11 |
+
"learning_rate": 1e-05,
|
| 12 |
+
"loss": 9.4473583984375,
|
| 13 |
+
"step": 1000
|
| 14 |
+
},
|
| 15 |
+
{
|
| 16 |
+
"epoch": 0.001479617163855024,
|
| 17 |
+
"learning_rate": 2e-05,
|
| 18 |
+
"loss": 4.761673828125,
|
| 19 |
+
"step": 2000
|
| 20 |
+
},
|
| 21 |
+
{
|
| 22 |
+
"epoch": 0.002219425745782536,
|
| 23 |
+
"learning_rate": 3e-05,
|
| 24 |
+
"loss": 3.3295283203125,
|
| 25 |
+
"step": 3000
|
| 26 |
+
},
|
| 27 |
+
{
|
| 28 |
+
"epoch": 0.002959234327710048,
|
| 29 |
+
"learning_rate": 4e-05,
|
| 30 |
+
"loss": 2.793072265625,
|
| 31 |
+
"step": 4000
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.00369904290963756,
|
| 35 |
+
"learning_rate": 5e-05,
|
| 36 |
+
"loss": 2.51521875,
|
| 37 |
+
"step": 5000
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.004438851491565072,
|
| 41 |
+
"learning_rate": 6e-05,
|
| 42 |
+
"loss": 2.3833671875,
|
| 43 |
+
"step": 6000
|
| 44 |
+
},
|
| 45 |
+
{
|
| 46 |
+
"epoch": 0.005178660073492584,
|
| 47 |
+
"learning_rate": 7e-05,
|
| 48 |
+
"loss": 2.28459765625,
|
| 49 |
+
"step": 7000
|
| 50 |
+
},
|
| 51 |
+
{
|
| 52 |
+
"epoch": 0.005918468655420096,
|
| 53 |
+
"learning_rate": 8e-05,
|
| 54 |
+
"loss": 2.228125,
|
| 55 |
+
"step": 8000
|
| 56 |
+
},
|
| 57 |
+
{
|
| 58 |
+
"epoch": 0.006658277237347608,
|
| 59 |
+
"learning_rate": 9e-05,
|
| 60 |
+
"loss": 2.19816015625,
|
| 61 |
+
"step": 9000
|
| 62 |
+
},
|
| 63 |
+
{
|
| 64 |
+
"epoch": 0.00739808581927512,
|
| 65 |
+
"learning_rate": 0.0001,
|
| 66 |
+
"loss": 2.15448828125,
|
| 67 |
+
"step": 10000
|
| 68 |
+
},
|
| 69 |
+
{
|
| 70 |
+
"epoch": 0.008137894401202633,
|
| 71 |
+
"learning_rate": 9.98989898989899e-05,
|
| 72 |
+
"loss": 2.13649609375,
|
| 73 |
+
"step": 11000
|
| 74 |
+
},
|
| 75 |
+
{
|
| 76 |
+
"epoch": 0.008877702983130144,
|
| 77 |
+
"learning_rate": 9.97979797979798e-05,
|
| 78 |
+
"loss": 2.08796484375,
|
| 79 |
+
"step": 12000
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.009617511565057657,
|
| 83 |
+
"learning_rate": 9.96969696969697e-05,
|
| 84 |
+
"loss": 2.0968984375,
|
| 85 |
+
"step": 13000
|
| 86 |
+
},
|
| 87 |
+
{
|
| 88 |
+
"epoch": 0.010357320146985169,
|
| 89 |
+
"learning_rate": 9.95959595959596e-05,
|
| 90 |
+
"loss": 2.08955078125,
|
| 91 |
+
"step": 14000
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 0.011097128728912681,
|
| 95 |
+
"learning_rate": 9.94949494949495e-05,
|
| 96 |
+
"loss": 2.0763828125,
|
| 97 |
+
"step": 15000
|
| 98 |
+
},
|
| 99 |
+
{
|
| 100 |
+
"epoch": 0.011836937310840193,
|
| 101 |
+
"learning_rate": 9.939393939393939e-05,
|
| 102 |
+
"loss": 2.06798828125,
|
| 103 |
+
"step": 16000
|
| 104 |
+
},
|
| 105 |
+
{
|
| 106 |
+
"epoch": 0.012576745892767706,
|
| 107 |
+
"learning_rate": 9.92929292929293e-05,
|
| 108 |
+
"loss": 2.06260546875,
|
| 109 |
+
"step": 17000
|
| 110 |
+
},
|
| 111 |
+
{
|
| 112 |
+
"epoch": 0.013316554474695217,
|
| 113 |
+
"learning_rate": 9.919191919191919e-05,
|
| 114 |
+
"loss": 2.058328125,
|
| 115 |
+
"step": 18000
|
| 116 |
+
},
|
| 117 |
+
{
|
| 118 |
+
"epoch": 0.01405636305662273,
|
| 119 |
+
"learning_rate": 9.909090909090911e-05,
|
| 120 |
+
"loss": 2.04747265625,
|
| 121 |
+
"step": 19000
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.01479617163855024,
|
| 125 |
+
"learning_rate": 9.8989898989899e-05,
|
| 126 |
+
"loss": 2.0418671875,
|
| 127 |
+
"step": 20000
|
| 128 |
+
},
|
| 129 |
+
{
|
| 130 |
+
"epoch": 0.015535980220477754,
|
| 131 |
+
"learning_rate": 9.888888888888889e-05,
|
| 132 |
+
"loss": 2.0486484375,
|
| 133 |
+
"step": 21000
|
| 134 |
+
},
|
| 135 |
+
{
|
| 136 |
+
"epoch": 0.016275788802405267,
|
| 137 |
+
"learning_rate": 9.87878787878788e-05,
|
| 138 |
+
"loss": 2.052984375,
|
| 139 |
+
"step": 22000
|
| 140 |
+
},
|
| 141 |
+
{
|
| 142 |
+
"epoch": 0.017015597384332778,
|
| 143 |
+
"learning_rate": 9.868686868686869e-05,
|
| 144 |
+
"loss": 2.0374609375,
|
| 145 |
+
"step": 23000
|
| 146 |
+
},
|
| 147 |
+
{
|
| 148 |
+
"epoch": 0.01775540596626029,
|
| 149 |
+
"learning_rate": 9.85858585858586e-05,
|
| 150 |
+
"loss": 2.03405078125,
|
| 151 |
+
"step": 24000
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 0.018495214548187804,
|
| 155 |
+
"learning_rate": 9.848484848484849e-05,
|
| 156 |
+
"loss": 2.04113671875,
|
| 157 |
+
"step": 25000
|
| 158 |
+
},
|
| 159 |
+
{
|
| 160 |
+
"epoch": 0.019235023130115315,
|
| 161 |
+
"learning_rate": 9.838383838383838e-05,
|
| 162 |
+
"loss": 2.02526953125,
|
| 163 |
+
"step": 26000
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.019974831712042826,
|
| 167 |
+
"learning_rate": 9.828282828282829e-05,
|
| 168 |
+
"loss": 2.0195234375,
|
| 169 |
+
"step": 27000
|
| 170 |
+
},
|
| 171 |
+
{
|
| 172 |
+
"epoch": 0.020714640293970337,
|
| 173 |
+
"learning_rate": 9.818181818181818e-05,
|
| 174 |
+
"loss": 2.0342734375,
|
| 175 |
+
"step": 28000
|
| 176 |
+
},
|
| 177 |
+
{
|
| 178 |
+
"epoch": 0.021454448875897852,
|
| 179 |
+
"learning_rate": 9.808080808080809e-05,
|
| 180 |
+
"loss": 2.0206328125,
|
| 181 |
+
"step": 29000
|
| 182 |
+
},
|
| 183 |
+
{
|
| 184 |
+
"epoch": 0.022194257457825363,
|
| 185 |
+
"learning_rate": 9.797979797979798e-05,
|
| 186 |
+
"loss": 2.0215859375,
|
| 187 |
+
"step": 30000
|
| 188 |
+
},
|
| 189 |
+
{
|
| 190 |
+
"epoch": 0.022934066039752874,
|
| 191 |
+
"learning_rate": 9.787878787878789e-05,
|
| 192 |
+
"loss": 2.0064453125,
|
| 193 |
+
"step": 31000
|
| 194 |
+
},
|
| 195 |
+
{
|
| 196 |
+
"epoch": 0.023673874621680385,
|
| 197 |
+
"learning_rate": 9.777777777777778e-05,
|
| 198 |
+
"loss": 2.0098984375,
|
| 199 |
+
"step": 32000
|
| 200 |
+
},
|
| 201 |
+
{
|
| 202 |
+
"epoch": 0.0244136832036079,
|
| 203 |
+
"learning_rate": 9.767676767676767e-05,
|
| 204 |
+
"loss": 2.0178046875,
|
| 205 |
+
"step": 33000
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.02515349178553541,
|
| 209 |
+
"learning_rate": 9.757575757575758e-05,
|
| 210 |
+
"loss": 2.01334375,
|
| 211 |
+
"step": 34000
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 0.025893300367462922,
|
| 215 |
+
"learning_rate": 9.747474747474747e-05,
|
| 216 |
+
"loss": 2.01625,
|
| 217 |
+
"step": 35000
|
| 218 |
+
},
|
| 219 |
+
{
|
| 220 |
+
"epoch": 0.026633108949390433,
|
| 221 |
+
"learning_rate": 9.737373737373738e-05,
|
| 222 |
+
"loss": 2.0108828125,
|
| 223 |
+
"step": 36000
|
| 224 |
+
},
|
| 225 |
+
{
|
| 226 |
+
"epoch": 0.027372917531317948,
|
| 227 |
+
"learning_rate": 9.727272727272728e-05,
|
| 228 |
+
"loss": 2.018734375,
|
| 229 |
+
"step": 37000
|
| 230 |
+
},
|
| 231 |
+
{
|
| 232 |
+
"epoch": 0.02811272611324546,
|
| 233 |
+
"learning_rate": 9.717171717171718e-05,
|
| 234 |
+
"loss": 2.003578125,
|
| 235 |
+
"step": 38000
|
| 236 |
+
},
|
| 237 |
+
{
|
| 238 |
+
"epoch": 0.02885253469517297,
|
| 239 |
+
"learning_rate": 9.707070707070708e-05,
|
| 240 |
+
"loss": 2.01503125,
|
| 241 |
+
"step": 39000
|
| 242 |
+
},
|
| 243 |
+
{
|
| 244 |
+
"epoch": 0.02959234327710048,
|
| 245 |
+
"learning_rate": 9.696969696969698e-05,
|
| 246 |
+
"loss": 2.016765625,
|
| 247 |
+
"step": 40000
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.030332151859027996,
|
| 251 |
+
"learning_rate": 9.686868686868688e-05,
|
| 252 |
+
"loss": 2.0055703125,
|
| 253 |
+
"step": 41000
|
| 254 |
+
},
|
| 255 |
+
{
|
| 256 |
+
"epoch": 0.031071960440955507,
|
| 257 |
+
"learning_rate": 9.676767676767677e-05,
|
| 258 |
+
"loss": 2.015375,
|
| 259 |
+
"step": 42000
|
| 260 |
+
},
|
| 261 |
+
{
|
| 262 |
+
"epoch": 0.03181176902288302,
|
| 263 |
+
"learning_rate": 9.666666666666667e-05,
|
| 264 |
+
"loss": 2.0087421875,
|
| 265 |
+
"step": 43000
|
| 266 |
+
},
|
| 267 |
+
{
|
| 268 |
+
"epoch": 0.03255157760481053,
|
| 269 |
+
"learning_rate": 9.656565656565657e-05,
|
| 270 |
+
"loss": 2.016359375,
|
| 271 |
+
"step": 44000
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 0.03329138618673804,
|
| 275 |
+
"learning_rate": 9.646464646464647e-05,
|
| 276 |
+
"loss": 2.015015625,
|
| 277 |
+
"step": 45000
|
| 278 |
+
},
|
| 279 |
+
{
|
| 280 |
+
"epoch": 0.034031194768665556,
|
| 281 |
+
"learning_rate": 9.636363636363637e-05,
|
| 282 |
+
"loss": 2.025421875,
|
| 283 |
+
"step": 46000
|
| 284 |
+
},
|
| 285 |
+
{
|
| 286 |
+
"epoch": 0.03477100335059307,
|
| 287 |
+
"learning_rate": 9.626262626262627e-05,
|
| 288 |
+
"loss": 2.0208671875,
|
| 289 |
+
"step": 47000
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.03551081193252058,
|
| 293 |
+
"learning_rate": 9.616161616161616e-05,
|
| 294 |
+
"loss": 2.00615625,
|
| 295 |
+
"step": 48000
|
| 296 |
+
},
|
| 297 |
+
{
|
| 298 |
+
"epoch": 0.03625062051444809,
|
| 299 |
+
"learning_rate": 9.606060606060606e-05,
|
| 300 |
+
"loss": 2.00828125,
|
| 301 |
+
"step": 49000
|
| 302 |
+
},
|
| 303 |
+
{
|
| 304 |
+
"epoch": 0.03699042909637561,
|
| 305 |
+
"learning_rate": 9.595959595959596e-05,
|
| 306 |
+
"loss": 1.998234375,
|
| 307 |
+
"step": 50000
|
| 308 |
+
},
|
| 309 |
+
{
|
| 310 |
+
"epoch": 0.037730237678303115,
|
| 311 |
+
"learning_rate": 9.585858585858586e-05,
|
| 312 |
+
"loss": 2.0086640625,
|
| 313 |
+
"step": 51000
|
| 314 |
+
},
|
| 315 |
+
{
|
| 316 |
+
"epoch": 0.03847004626023063,
|
| 317 |
+
"learning_rate": 9.575757575757576e-05,
|
| 318 |
+
"loss": 1.9955234375,
|
| 319 |
+
"step": 52000
|
| 320 |
+
},
|
| 321 |
+
{
|
| 322 |
+
"epoch": 0.03920985484215814,
|
| 323 |
+
"learning_rate": 9.565656565656566e-05,
|
| 324 |
+
"loss": 2.0098984375,
|
| 325 |
+
"step": 53000
|
| 326 |
+
},
|
| 327 |
+
{
|
| 328 |
+
"epoch": 0.03994966342408565,
|
| 329 |
+
"learning_rate": 9.555555555555557e-05,
|
| 330 |
+
"loss": 1.9921015625,
|
| 331 |
+
"step": 54000
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.040689472006013166,
|
| 335 |
+
"learning_rate": 9.545454545454546e-05,
|
| 336 |
+
"loss": 2.0024765625,
|
| 337 |
+
"step": 55000
|
| 338 |
+
},
|
| 339 |
+
{
|
| 340 |
+
"epoch": 0.041429280587940674,
|
| 341 |
+
"learning_rate": 9.535353535353537e-05,
|
| 342 |
+
"loss": 2.0018515625,
|
| 343 |
+
"step": 56000
|
| 344 |
+
},
|
| 345 |
+
{
|
| 346 |
+
"epoch": 0.04216908916986819,
|
| 347 |
+
"learning_rate": 9.525252525252526e-05,
|
| 348 |
+
"loss": 2.009984375,
|
| 349 |
+
"step": 57000
|
| 350 |
+
},
|
| 351 |
+
{
|
| 352 |
+
"epoch": 0.042908897751795703,
|
| 353 |
+
"learning_rate": 9.515151515151515e-05,
|
| 354 |
+
"loss": 2.00109375,
|
| 355 |
+
"step": 58000
|
| 356 |
+
},
|
| 357 |
+
{
|
| 358 |
+
"epoch": 0.04364870633372321,
|
| 359 |
+
"learning_rate": 9.505050505050506e-05,
|
| 360 |
+
"loss": 2.017796875,
|
| 361 |
+
"step": 59000
|
| 362 |
+
},
|
| 363 |
+
{
|
| 364 |
+
"epoch": 0.044388514915650726,
|
| 365 |
+
"learning_rate": 9.494949494949495e-05,
|
| 366 |
+
"loss": 2.005921875,
|
| 367 |
+
"step": 60000
|
| 368 |
+
},
|
| 369 |
+
{
|
| 370 |
+
"epoch": 0.045128323497578234,
|
| 371 |
+
"learning_rate": 9.484848484848486e-05,
|
| 372 |
+
"loss": 2.012265625,
|
| 373 |
+
"step": 61000
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.04586813207950575,
|
| 377 |
+
"learning_rate": 9.474747474747475e-05,
|
| 378 |
+
"loss": 1.99825,
|
| 379 |
+
"step": 62000
|
| 380 |
+
},
|
| 381 |
+
{
|
| 382 |
+
"epoch": 0.04660794066143326,
|
| 383 |
+
"learning_rate": 9.464646464646464e-05,
|
| 384 |
+
"loss": 2.003953125,
|
| 385 |
+
"step": 63000
|
| 386 |
+
},
|
| 387 |
+
{
|
| 388 |
+
"epoch": 0.04734774924336077,
|
| 389 |
+
"learning_rate": 9.454545454545455e-05,
|
| 390 |
+
"loss": 2.019515625,
|
| 391 |
+
"step": 64000
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 0.048087557825288285,
|
| 395 |
+
"learning_rate": 9.444444444444444e-05,
|
| 396 |
+
"loss": 2.00675,
|
| 397 |
+
"step": 65000
|
| 398 |
+
},
|
| 399 |
+
{
|
| 400 |
+
"epoch": 0.0488273664072158,
|
| 401 |
+
"learning_rate": 9.434343434343435e-05,
|
| 402 |
+
"loss": 2.005234375,
|
| 403 |
+
"step": 66000
|
| 404 |
+
},
|
| 405 |
+
{
|
| 406 |
+
"epoch": 0.04956717498914331,
|
| 407 |
+
"learning_rate": 9.424242424242424e-05,
|
| 408 |
+
"loss": 1.9940625,
|
| 409 |
+
"step": 67000
|
| 410 |
+
},
|
| 411 |
+
{
|
| 412 |
+
"epoch": 0.05030698357107082,
|
| 413 |
+
"learning_rate": 9.414141414141415e-05,
|
| 414 |
+
"loss": 1.999109375,
|
| 415 |
+
"step": 68000
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 0.05104679215299833,
|
| 419 |
+
"learning_rate": 9.404040404040404e-05,
|
| 420 |
+
"loss": 2.00134375,
|
| 421 |
+
"step": 69000
|
| 422 |
+
},
|
| 423 |
+
{
|
| 424 |
+
"epoch": 0.051786600734925844,
|
| 425 |
+
"learning_rate": 9.393939393939395e-05,
|
| 426 |
+
"loss": 2.007953125,
|
| 427 |
+
"step": 70000
|
| 428 |
+
},
|
| 429 |
+
{
|
| 430 |
+
"epoch": 0.05252640931685336,
|
| 431 |
+
"learning_rate": 9.383838383838385e-05,
|
| 432 |
+
"loss": 2.005171875,
|
| 433 |
+
"step": 71000
|
| 434 |
+
},
|
| 435 |
+
{
|
| 436 |
+
"epoch": 0.05326621789878087,
|
| 437 |
+
"learning_rate": 9.373737373737375e-05,
|
| 438 |
+
"loss": 2.000734375,
|
| 439 |
+
"step": 72000
|
| 440 |
+
},
|
| 441 |
+
{
|
| 442 |
+
"epoch": 0.05400602648070838,
|
| 443 |
+
"learning_rate": 9.363636363636364e-05,
|
| 444 |
+
"loss": 2.01009375,
|
| 445 |
+
"step": 73000
|
| 446 |
+
},
|
| 447 |
+
{
|
| 448 |
+
"epoch": 0.054745835062635896,
|
| 449 |
+
"learning_rate": 9.353535353535354e-05,
|
| 450 |
+
"loss": 2.006703125,
|
| 451 |
+
"step": 74000
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 0.055485643644563404,
|
| 455 |
+
"learning_rate": 9.343434343434344e-05,
|
| 456 |
+
"loss": 2.019640625,
|
| 457 |
+
"step": 75000
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 0.05622545222649092,
|
| 461 |
+
"learning_rate": 9.333333333333334e-05,
|
| 462 |
+
"loss": 1.999078125,
|
| 463 |
+
"step": 76000
|
| 464 |
+
},
|
| 465 |
+
{
|
| 466 |
+
"epoch": 0.05696526080841843,
|
| 467 |
+
"learning_rate": 9.323232323232324e-05,
|
| 468 |
+
"loss": 1.995890625,
|
| 469 |
+
"step": 77000
|
| 470 |
+
},
|
| 471 |
+
{
|
| 472 |
+
"epoch": 0.05770506939034594,
|
| 473 |
+
"learning_rate": 9.313131313131314e-05,
|
| 474 |
+
"loss": 2.00525,
|
| 475 |
+
"step": 78000
|
| 476 |
+
},
|
| 477 |
+
{
|
| 478 |
+
"epoch": 0.058444877972273455,
|
| 479 |
+
"learning_rate": 9.303030303030303e-05,
|
| 480 |
+
"loss": 2.0128125,
|
| 481 |
+
"step": 79000
|
| 482 |
+
},
|
| 483 |
+
{
|
| 484 |
+
"epoch": 0.05918468655420096,
|
| 485 |
+
"learning_rate": 9.292929292929293e-05,
|
| 486 |
+
"loss": 2.006140625,
|
| 487 |
+
"step": 80000
|
| 488 |
+
},
|
| 489 |
+
{
|
| 490 |
+
"epoch": 0.05992449513612848,
|
| 491 |
+
"learning_rate": 9.282828282828283e-05,
|
| 492 |
+
"loss": 2.0091875,
|
| 493 |
+
"step": 81000
|
| 494 |
+
},
|
| 495 |
+
{
|
| 496 |
+
"epoch": 0.06066430371805599,
|
| 497 |
+
"learning_rate": 9.272727272727273e-05,
|
| 498 |
+
"loss": 1.997203125,
|
| 499 |
+
"step": 82000
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 0.0614041122999835,
|
| 503 |
+
"learning_rate": 9.262626262626263e-05,
|
| 504 |
+
"loss": 2.020015625,
|
| 505 |
+
"step": 83000
|
| 506 |
+
},
|
| 507 |
+
{
|
| 508 |
+
"epoch": 0.062143920881911015,
|
| 509 |
+
"learning_rate": 9.252525252525253e-05,
|
| 510 |
+
"loss": 2.028375,
|
| 511 |
+
"step": 84000
|
| 512 |
+
},
|
| 513 |
+
{
|
| 514 |
+
"epoch": 0.06288372946383852,
|
| 515 |
+
"learning_rate": 9.242424242424242e-05,
|
| 516 |
+
"loss": 2.006765625,
|
| 517 |
+
"step": 85000
|
| 518 |
+
},
|
| 519 |
+
{
|
| 520 |
+
"epoch": 0.06362353804576604,
|
| 521 |
+
"learning_rate": 9.232323232323232e-05,
|
| 522 |
+
"loss": 2.016703125,
|
| 523 |
+
"step": 86000
|
| 524 |
+
},
|
| 525 |
+
{
|
| 526 |
+
"epoch": 0.06436334662769355,
|
| 527 |
+
"learning_rate": 9.222222222222223e-05,
|
| 528 |
+
"loss": 2.041,
|
| 529 |
+
"step": 87000
|
| 530 |
+
},
|
| 531 |
+
{
|
| 532 |
+
"epoch": 0.06510315520962107,
|
| 533 |
+
"learning_rate": 9.212121212121214e-05,
|
| 534 |
+
"loss": 2.029640625,
|
| 535 |
+
"step": 88000
|
| 536 |
+
},
|
| 537 |
+
{
|
| 538 |
+
"epoch": 0.06584296379154858,
|
| 539 |
+
"learning_rate": 9.202020202020203e-05,
|
| 540 |
+
"loss": 2.005890625,
|
| 541 |
+
"step": 89000
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 0.06658277237347608,
|
| 545 |
+
"learning_rate": 9.191919191919192e-05,
|
| 546 |
+
"loss": 1.997921875,
|
| 547 |
+
"step": 90000
|
| 548 |
+
},
|
| 549 |
+
{
|
| 550 |
+
"epoch": 0.0673225809554036,
|
| 551 |
+
"learning_rate": 9.181818181818183e-05,
|
| 552 |
+
"loss": 1.993796875,
|
| 553 |
+
"step": 91000
|
| 554 |
+
},
|
| 555 |
+
{
|
| 556 |
+
"epoch": 0.06806238953733111,
|
| 557 |
+
"learning_rate": 9.171717171717172e-05,
|
| 558 |
+
"loss": 1.997109375,
|
| 559 |
+
"step": 92000
|
| 560 |
+
},
|
| 561 |
+
{
|
| 562 |
+
"epoch": 0.06880219811925863,
|
| 563 |
+
"learning_rate": 9.161616161616163e-05,
|
| 564 |
+
"loss": 2.004078125,
|
| 565 |
+
"step": 93000
|
| 566 |
+
},
|
| 567 |
+
{
|
| 568 |
+
"epoch": 0.06954200670118614,
|
| 569 |
+
"learning_rate": 9.151515151515152e-05,
|
| 570 |
+
"loss": 2.001828125,
|
| 571 |
+
"step": 94000
|
| 572 |
+
},
|
| 573 |
+
{
|
| 574 |
+
"epoch": 0.07028181528311364,
|
| 575 |
+
"learning_rate": 9.141414141414141e-05,
|
| 576 |
+
"loss": 1.9931875,
|
| 577 |
+
"step": 95000
|
| 578 |
+
},
|
| 579 |
+
{
|
| 580 |
+
"epoch": 0.07102162386504116,
|
| 581 |
+
"learning_rate": 9.131313131313132e-05,
|
| 582 |
+
"loss": 1.987171875,
|
| 583 |
+
"step": 96000
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 0.07176143244696867,
|
| 587 |
+
"learning_rate": 9.121212121212121e-05,
|
| 588 |
+
"loss": 2.0305,
|
| 589 |
+
"step": 97000
|
| 590 |
+
},
|
| 591 |
+
{
|
| 592 |
+
"epoch": 0.07250124102889619,
|
| 593 |
+
"learning_rate": 9.111111111111112e-05,
|
| 594 |
+
"loss": 1.999484375,
|
| 595 |
+
"step": 98000
|
| 596 |
+
},
|
| 597 |
+
{
|
| 598 |
+
"epoch": 0.0732410496108237,
|
| 599 |
+
"learning_rate": 9.101010101010101e-05,
|
| 600 |
+
"loss": 1.989171875,
|
| 601 |
+
"step": 99000
|
| 602 |
+
},
|
| 603 |
+
{
|
| 604 |
+
"epoch": 0.07398085819275121,
|
| 605 |
+
"learning_rate": 9.090909090909092e-05,
|
| 606 |
+
"loss": 1.988671875,
|
| 607 |
+
"step": 100000
|
| 608 |
+
},
|
| 609 |
+
{
|
| 610 |
+
"epoch": 0.07472066677467872,
|
| 611 |
+
"learning_rate": 9.080808080808081e-05,
|
| 612 |
+
"loss": 1.989,
|
| 613 |
+
"step": 101000
|
| 614 |
+
},
|
| 615 |
+
{
|
| 616 |
+
"epoch": 0.07546047535660623,
|
| 617 |
+
"learning_rate": 9.07070707070707e-05,
|
| 618 |
+
"loss": 1.99684375,
|
| 619 |
+
"step": 102000
|
| 620 |
+
},
|
| 621 |
+
{
|
| 622 |
+
"epoch": 0.07620028393853374,
|
| 623 |
+
"learning_rate": 9.060606060606061e-05,
|
| 624 |
+
"loss": 2.004921875,
|
| 625 |
+
"step": 103000
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 0.07694009252046126,
|
| 629 |
+
"learning_rate": 9.050505050505052e-05,
|
| 630 |
+
"loss": 2.01390625,
|
| 631 |
+
"step": 104000
|
| 632 |
+
},
|
| 633 |
+
{
|
| 634 |
+
"epoch": 0.07767990110238877,
|
| 635 |
+
"learning_rate": 9.040404040404041e-05,
|
| 636 |
+
"loss": 1.98865625,
|
| 637 |
+
"step": 105000
|
| 638 |
+
},
|
| 639 |
+
{
|
| 640 |
+
"epoch": 0.07841970968431627,
|
| 641 |
+
"learning_rate": 9.030303030303031e-05,
|
| 642 |
+
"loss": 1.98871875,
|
| 643 |
+
"step": 106000
|
| 644 |
+
},
|
| 645 |
+
{
|
| 646 |
+
"epoch": 0.07915951826624379,
|
| 647 |
+
"learning_rate": 9.02020202020202e-05,
|
| 648 |
+
"loss": 1.994046875,
|
| 649 |
+
"step": 107000
|
| 650 |
+
},
|
| 651 |
+
{
|
| 652 |
+
"epoch": 0.0798993268481713,
|
| 653 |
+
"learning_rate": 9.010101010101011e-05,
|
| 654 |
+
"loss": 2.010546875,
|
| 655 |
+
"step": 108000
|
| 656 |
+
},
|
| 657 |
+
{
|
| 658 |
+
"epoch": 0.08063913543009882,
|
| 659 |
+
"learning_rate": 9e-05,
|
| 660 |
+
"loss": 2.005640625,
|
| 661 |
+
"step": 109000
|
| 662 |
+
},
|
| 663 |
+
{
|
| 664 |
+
"epoch": 0.08137894401202633,
|
| 665 |
+
"learning_rate": 8.98989898989899e-05,
|
| 666 |
+
"loss": 1.99709375,
|
| 667 |
+
"step": 110000
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 0.08211875259395383,
|
| 671 |
+
"learning_rate": 8.97979797979798e-05,
|
| 672 |
+
"loss": 2.01225,
|
| 673 |
+
"step": 111000
|
| 674 |
+
},
|
| 675 |
+
{
|
| 676 |
+
"epoch": 0.08285856117588135,
|
| 677 |
+
"learning_rate": 8.96969696969697e-05,
|
| 678 |
+
"loss": 1.982421875,
|
| 679 |
+
"step": 112000
|
| 680 |
+
},
|
| 681 |
+
{
|
| 682 |
+
"epoch": 0.08359836975780886,
|
| 683 |
+
"learning_rate": 8.95959595959596e-05,
|
| 684 |
+
"loss": 1.988484375,
|
| 685 |
+
"step": 113000
|
| 686 |
+
},
|
| 687 |
+
{
|
| 688 |
+
"epoch": 0.08433817833973638,
|
| 689 |
+
"learning_rate": 8.94949494949495e-05,
|
| 690 |
+
"loss": 2.001359375,
|
| 691 |
+
"step": 114000
|
| 692 |
+
},
|
| 693 |
+
{
|
| 694 |
+
"epoch": 0.08507798692166389,
|
| 695 |
+
"learning_rate": 8.93939393939394e-05,
|
| 696 |
+
"loss": 2.001796875,
|
| 697 |
+
"step": 115000
|
| 698 |
+
},
|
| 699 |
+
{
|
| 700 |
+
"epoch": 0.08581779550359141,
|
| 701 |
+
"learning_rate": 8.92929292929293e-05,
|
| 702 |
+
"loss": 2.012875,
|
| 703 |
+
"step": 116000
|
| 704 |
+
},
|
| 705 |
+
{
|
| 706 |
+
"epoch": 0.08655760408551891,
|
| 707 |
+
"learning_rate": 8.919191919191919e-05,
|
| 708 |
+
"loss": 2.007953125,
|
| 709 |
+
"step": 117000
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"epoch": 0.08729741266744642,
|
| 713 |
+
"learning_rate": 8.90909090909091e-05,
|
| 714 |
+
"loss": 2.00078125,
|
| 715 |
+
"step": 118000
|
| 716 |
+
},
|
| 717 |
+
{
|
| 718 |
+
"epoch": 0.08803722124937394,
|
| 719 |
+
"learning_rate": 8.898989898989899e-05,
|
| 720 |
+
"loss": 2.02653125,
|
| 721 |
+
"step": 119000
|
| 722 |
+
},
|
| 723 |
+
{
|
| 724 |
+
"epoch": 0.08877702983130145,
|
| 725 |
+
"learning_rate": 8.888888888888889e-05,
|
| 726 |
+
"loss": 2.006796875,
|
| 727 |
+
"step": 120000
|
| 728 |
+
},
|
| 729 |
+
{
|
| 730 |
+
"epoch": 0.08951683841322897,
|
| 731 |
+
"learning_rate": 8.87878787878788e-05,
|
| 732 |
+
"loss": 2.007296875,
|
| 733 |
+
"step": 121000
|
| 734 |
+
},
|
| 735 |
+
{
|
| 736 |
+
"epoch": 0.09025664699515647,
|
| 737 |
+
"learning_rate": 8.868686868686869e-05,
|
| 738 |
+
"loss": 2.012890625,
|
| 739 |
+
"step": 122000
|
| 740 |
+
},
|
| 741 |
+
{
|
| 742 |
+
"epoch": 0.09099645557708398,
|
| 743 |
+
"learning_rate": 8.85858585858586e-05,
|
| 744 |
+
"loss": 2.012546875,
|
| 745 |
+
"step": 123000
|
| 746 |
+
},
|
| 747 |
+
{
|
| 748 |
+
"epoch": 0.0917362641590115,
|
| 749 |
+
"learning_rate": 8.848484848484849e-05,
|
| 750 |
+
"loss": 2.016765625,
|
| 751 |
+
"step": 124000
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 0.09247607274093901,
|
| 755 |
+
"learning_rate": 8.83838383838384e-05,
|
| 756 |
+
"loss": 1.985875,
|
| 757 |
+
"step": 125000
|
| 758 |
+
},
|
| 759 |
+
{
|
| 760 |
+
"epoch": 0.09321588132286653,
|
| 761 |
+
"learning_rate": 8.828282828282829e-05,
|
| 762 |
+
"loss": 2.0315,
|
| 763 |
+
"step": 126000
|
| 764 |
+
},
|
| 765 |
+
{
|
| 766 |
+
"epoch": 0.09395568990479404,
|
| 767 |
+
"learning_rate": 8.818181818181818e-05,
|
| 768 |
+
"loss": 2.061625,
|
| 769 |
+
"step": 127000
|
| 770 |
+
},
|
| 771 |
+
{
|
| 772 |
+
"epoch": 0.09469549848672154,
|
| 773 |
+
"learning_rate": 8.808080808080809e-05,
|
| 774 |
+
"loss": 2.03778125,
|
| 775 |
+
"step": 128000
|
| 776 |
+
},
|
| 777 |
+
{
|
| 778 |
+
"epoch": 0.09543530706864906,
|
| 779 |
+
"learning_rate": 8.797979797979798e-05,
|
| 780 |
+
"loss": 2.0075625,
|
| 781 |
+
"step": 129000
|
| 782 |
+
},
|
| 783 |
+
{
|
| 784 |
+
"epoch": 0.09617511565057657,
|
| 785 |
+
"learning_rate": 8.787878787878789e-05,
|
| 786 |
+
"loss": 1.9984375,
|
| 787 |
+
"step": 130000
|
| 788 |
+
},
|
| 789 |
+
{
|
| 790 |
+
"epoch": 0.09691492423250408,
|
| 791 |
+
"learning_rate": 8.777777777777778e-05,
|
| 792 |
+
"loss": 1.99621875,
|
| 793 |
+
"step": 131000
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 0.0976547328144316,
|
| 797 |
+
"learning_rate": 8.767676767676767e-05,
|
| 798 |
+
"loss": 2.01059375,
|
| 799 |
+
"step": 132000
|
| 800 |
+
},
|
| 801 |
+
{
|
| 802 |
+
"epoch": 0.0983945413963591,
|
| 803 |
+
"learning_rate": 8.757575757575758e-05,
|
| 804 |
+
"loss": 2.0471875,
|
| 805 |
+
"step": 133000
|
| 806 |
+
},
|
| 807 |
+
{
|
| 808 |
+
"epoch": 0.09913434997828661,
|
| 809 |
+
"learning_rate": 8.747474747474747e-05,
|
| 810 |
+
"loss": 2.01190625,
|
| 811 |
+
"step": 134000
|
| 812 |
+
},
|
| 813 |
+
{
|
| 814 |
+
"epoch": 0.09987415856021413,
|
| 815 |
+
"learning_rate": 8.737373737373738e-05,
|
| 816 |
+
"loss": 2.00653125,
|
| 817 |
+
"step": 135000
|
| 818 |
+
},
|
| 819 |
+
{
|
| 820 |
+
"epoch": 0.10061396714214164,
|
| 821 |
+
"learning_rate": 8.727272727272727e-05,
|
| 822 |
+
"loss": 2.014375,
|
| 823 |
+
"step": 136000
|
| 824 |
+
},
|
| 825 |
+
{
|
| 826 |
+
"epoch": 0.10135377572406916,
|
| 827 |
+
"learning_rate": 8.717171717171718e-05,
|
| 828 |
+
"loss": 2.00065625,
|
| 829 |
+
"step": 137000
|
| 830 |
+
},
|
| 831 |
+
{
|
| 832 |
+
"epoch": 0.10209358430599666,
|
| 833 |
+
"learning_rate": 8.707070707070707e-05,
|
| 834 |
+
"loss": 2.031875,
|
| 835 |
+
"step": 138000
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"epoch": 0.10283339288792417,
|
| 839 |
+
"learning_rate": 8.696969696969698e-05,
|
| 840 |
+
"loss": 1.99303125,
|
| 841 |
+
"step": 139000
|
| 842 |
+
},
|
| 843 |
+
{
|
| 844 |
+
"epoch": 0.10357320146985169,
|
| 845 |
+
"learning_rate": 8.686868686868688e-05,
|
| 846 |
+
"loss": 1.97703125,
|
| 847 |
+
"step": 140000
|
| 848 |
+
},
|
| 849 |
+
{
|
| 850 |
+
"epoch": 0.1043130100517792,
|
| 851 |
+
"learning_rate": 8.676767676767678e-05,
|
| 852 |
+
"loss": 1.99440625,
|
| 853 |
+
"step": 141000
|
| 854 |
+
},
|
| 855 |
+
{
|
| 856 |
+
"epoch": 0.10505281863370672,
|
| 857 |
+
"learning_rate": 8.666666666666667e-05,
|
| 858 |
+
"loss": 2.00790625,
|
| 859 |
+
"step": 142000
|
| 860 |
+
},
|
| 861 |
+
{
|
| 862 |
+
"epoch": 0.10579262721563423,
|
| 863 |
+
"learning_rate": 8.656565656565657e-05,
|
| 864 |
+
"loss": 1.9769375,
|
| 865 |
+
"step": 143000
|
| 866 |
+
},
|
| 867 |
+
{
|
| 868 |
+
"epoch": 0.10653243579756173,
|
| 869 |
+
"learning_rate": 8.646464646464647e-05,
|
| 870 |
+
"loss": 1.98775,
|
| 871 |
+
"step": 144000
|
| 872 |
+
},
|
| 873 |
+
{
|
| 874 |
+
"epoch": 0.10727224437948925,
|
| 875 |
+
"learning_rate": 8.636363636363637e-05,
|
| 876 |
+
"loss": 1.970875,
|
| 877 |
+
"step": 145000
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"epoch": 0.10801205296141676,
|
| 881 |
+
"learning_rate": 8.626262626262627e-05,
|
| 882 |
+
"loss": 1.97203125,
|
| 883 |
+
"step": 146000
|
| 884 |
+
},
|
| 885 |
+
{
|
| 886 |
+
"epoch": 0.10875186154334428,
|
| 887 |
+
"learning_rate": 8.616161616161616e-05,
|
| 888 |
+
"loss": 1.97359375,
|
| 889 |
+
"step": 147000
|
| 890 |
+
},
|
| 891 |
+
{
|
| 892 |
+
"epoch": 0.10949167012527179,
|
| 893 |
+
"learning_rate": 8.606060606060606e-05,
|
| 894 |
+
"loss": 1.97471875,
|
| 895 |
+
"step": 148000
|
| 896 |
+
},
|
| 897 |
+
{
|
| 898 |
+
"epoch": 0.11023147870719929,
|
| 899 |
+
"learning_rate": 8.595959595959596e-05,
|
| 900 |
+
"loss": 1.96978125,
|
| 901 |
+
"step": 149000
|
| 902 |
+
},
|
| 903 |
+
{
|
| 904 |
+
"epoch": 0.11097128728912681,
|
| 905 |
+
"learning_rate": 8.585858585858586e-05,
|
| 906 |
+
"loss": 1.9813125,
|
| 907 |
+
"step": 150000
|
| 908 |
+
},
|
| 909 |
+
{
|
| 910 |
+
"epoch": 0.11171109587105432,
|
| 911 |
+
"learning_rate": 8.575757575757576e-05,
|
| 912 |
+
"loss": 1.9761875,
|
| 913 |
+
"step": 151000
|
| 914 |
+
},
|
| 915 |
+
{
|
| 916 |
+
"epoch": 0.11245090445298184,
|
| 917 |
+
"learning_rate": 8.565656565656566e-05,
|
| 918 |
+
"loss": 1.9814375,
|
| 919 |
+
"step": 152000
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"epoch": 0.11319071303490935,
|
| 923 |
+
"learning_rate": 8.555555555555556e-05,
|
| 924 |
+
"loss": 1.97459375,
|
| 925 |
+
"step": 153000
|
| 926 |
+
},
|
| 927 |
+
{
|
| 928 |
+
"epoch": 0.11393052161683687,
|
| 929 |
+
"learning_rate": 8.545454545454545e-05,
|
| 930 |
+
"loss": 1.97503125,
|
| 931 |
+
"step": 154000
|
| 932 |
+
},
|
| 933 |
+
{
|
| 934 |
+
"epoch": 0.11467033019876437,
|
| 935 |
+
"learning_rate": 8.535353535353535e-05,
|
| 936 |
+
"loss": 1.98446875,
|
| 937 |
+
"step": 155000
|
| 938 |
+
},
|
| 939 |
+
{
|
| 940 |
+
"epoch": 0.11541013878069188,
|
| 941 |
+
"learning_rate": 8.525252525252526e-05,
|
| 942 |
+
"loss": 1.997,
|
| 943 |
+
"step": 156000
|
| 944 |
+
},
|
| 945 |
+
{
|
| 946 |
+
"epoch": 0.1161499473626194,
|
| 947 |
+
"learning_rate": 8.515151515151515e-05,
|
| 948 |
+
"loss": 1.98559375,
|
| 949 |
+
"step": 157000
|
| 950 |
+
},
|
| 951 |
+
{
|
| 952 |
+
"epoch": 0.11688975594454691,
|
| 953 |
+
"learning_rate": 8.505050505050506e-05,
|
| 954 |
+
"loss": 1.99725,
|
| 955 |
+
"step": 158000
|
| 956 |
+
},
|
| 957 |
+
{
|
| 958 |
+
"epoch": 0.11762956452647443,
|
| 959 |
+
"learning_rate": 8.494949494949495e-05,
|
| 960 |
+
"loss": 1.997375,
|
| 961 |
+
"step": 159000
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"epoch": 0.11836937310840193,
|
| 965 |
+
"learning_rate": 8.484848484848486e-05,
|
| 966 |
+
"loss": 1.98828125,
|
| 967 |
+
"step": 160000
|
| 968 |
+
},
|
| 969 |
+
{
|
| 970 |
+
"epoch": 0.11910918169032944,
|
| 971 |
+
"learning_rate": 8.474747474747475e-05,
|
| 972 |
+
"loss": 1.9948125,
|
| 973 |
+
"step": 161000
|
| 974 |
+
},
|
| 975 |
+
{
|
| 976 |
+
"epoch": 0.11984899027225696,
|
| 977 |
+
"learning_rate": 8.464646464646466e-05,
|
| 978 |
+
"loss": 1.986,
|
| 979 |
+
"step": 162000
|
| 980 |
+
},
|
| 981 |
+
{
|
| 982 |
+
"epoch": 0.12058879885418447,
|
| 983 |
+
"learning_rate": 8.454545454545455e-05,
|
| 984 |
+
"loss": 1.99940625,
|
| 985 |
+
"step": 163000
|
| 986 |
+
},
|
| 987 |
+
{
|
| 988 |
+
"epoch": 0.12132860743611198,
|
| 989 |
+
"learning_rate": 8.444444444444444e-05,
|
| 990 |
+
"loss": 1.99140625,
|
| 991 |
+
"step": 164000
|
| 992 |
+
},
|
| 993 |
+
{
|
| 994 |
+
"epoch": 0.12206841601803949,
|
| 995 |
+
"learning_rate": 8.434343434343435e-05,
|
| 996 |
+
"loss": 1.996375,
|
| 997 |
+
"step": 165000
|
| 998 |
+
},
|
| 999 |
+
{
|
| 1000 |
+
"epoch": 0.122808224599967,
|
| 1001 |
+
"learning_rate": 8.424242424242424e-05,
|
| 1002 |
+
"loss": 1.99515625,
|
| 1003 |
+
"step": 166000
|
| 1004 |
+
},
|
| 1005 |
+
{
|
| 1006 |
+
"epoch": 0.12354803318189451,
|
| 1007 |
+
"learning_rate": 8.414141414141415e-05,
|
| 1008 |
+
"loss": 1.97484375,
|
| 1009 |
+
"step": 167000
|
| 1010 |
+
},
|
| 1011 |
+
{
|
| 1012 |
+
"epoch": 0.12428784176382203,
|
| 1013 |
+
"learning_rate": 8.404040404040404e-05,
|
| 1014 |
+
"loss": 1.99240625,
|
| 1015 |
+
"step": 168000
|
| 1016 |
+
},
|
| 1017 |
+
{
|
| 1018 |
+
"epoch": 0.12502765034574953,
|
| 1019 |
+
"learning_rate": 8.393939393939393e-05,
|
| 1020 |
+
"loss": 1.99184375,
|
| 1021 |
+
"step": 169000
|
| 1022 |
+
},
|
| 1023 |
+
{
|
| 1024 |
+
"epoch": 0.12576745892767704,
|
| 1025 |
+
"learning_rate": 8.383838383838384e-05,
|
| 1026 |
+
"loss": 1.99071875,
|
| 1027 |
+
"step": 170000
|
| 1028 |
+
},
|
| 1029 |
+
{
|
| 1030 |
+
"epoch": 0.12650726750960456,
|
| 1031 |
+
"learning_rate": 8.373737373737373e-05,
|
| 1032 |
+
"loss": 1.99640625,
|
| 1033 |
+
"step": 171000
|
| 1034 |
+
},
|
| 1035 |
+
{
|
| 1036 |
+
"epoch": 0.12724707609153207,
|
| 1037 |
+
"learning_rate": 8.363636363636364e-05,
|
| 1038 |
+
"loss": 2.009375,
|
| 1039 |
+
"step": 172000
|
| 1040 |
+
},
|
| 1041 |
+
{
|
| 1042 |
+
"epoch": 0.1279868846734596,
|
| 1043 |
+
"learning_rate": 8.353535353535355e-05,
|
| 1044 |
+
"loss": 2.01953125,
|
| 1045 |
+
"step": 173000
|
| 1046 |
+
},
|
| 1047 |
+
{
|
| 1048 |
+
"epoch": 0.1287266932553871,
|
| 1049 |
+
"learning_rate": 8.343434343434344e-05,
|
| 1050 |
+
"loss": 2.0111875,
|
| 1051 |
+
"step": 174000
|
| 1052 |
+
},
|
| 1053 |
+
{
|
| 1054 |
+
"epoch": 0.12946650183731462,
|
| 1055 |
+
"learning_rate": 8.333333333333334e-05,
|
| 1056 |
+
"loss": 2.014,
|
| 1057 |
+
"step": 175000
|
| 1058 |
+
},
|
| 1059 |
+
{
|
| 1060 |
+
"epoch": 0.13020631041924213,
|
| 1061 |
+
"learning_rate": 8.323232323232324e-05,
|
| 1062 |
+
"loss": 2.0209375,
|
| 1063 |
+
"step": 176000
|
| 1064 |
+
},
|
| 1065 |
+
{
|
| 1066 |
+
"epoch": 0.13094611900116965,
|
| 1067 |
+
"learning_rate": 8.313131313131314e-05,
|
| 1068 |
+
"loss": 1.98971875,
|
| 1069 |
+
"step": 177000
|
| 1070 |
+
},
|
| 1071 |
+
{
|
| 1072 |
+
"epoch": 0.13168592758309716,
|
| 1073 |
+
"learning_rate": 8.303030303030304e-05,
|
| 1074 |
+
"loss": 1.978,
|
| 1075 |
+
"step": 178000
|
| 1076 |
+
},
|
| 1077 |
+
{
|
| 1078 |
+
"epoch": 0.13242573616502465,
|
| 1079 |
+
"learning_rate": 8.292929292929293e-05,
|
| 1080 |
+
"loss": 2.0005,
|
| 1081 |
+
"step": 179000
|
| 1082 |
+
},
|
| 1083 |
+
{
|
| 1084 |
+
"epoch": 0.13316554474695216,
|
| 1085 |
+
"learning_rate": 8.282828282828283e-05,
|
| 1086 |
+
"loss": 2.01215625,
|
| 1087 |
+
"step": 180000
|
| 1088 |
+
},
|
| 1089 |
+
{
|
| 1090 |
+
"epoch": 0.13390535332887968,
|
| 1091 |
+
"learning_rate": 8.272727272727273e-05,
|
| 1092 |
+
"loss": 2.0359375,
|
| 1093 |
+
"step": 181000
|
| 1094 |
+
},
|
| 1095 |
+
{
|
| 1096 |
+
"epoch": 0.1346451619108072,
|
| 1097 |
+
"learning_rate": 8.262626262626263e-05,
|
| 1098 |
+
"loss": 2.0521875,
|
| 1099 |
+
"step": 182000
|
| 1100 |
+
},
|
| 1101 |
+
{
|
| 1102 |
+
"epoch": 0.1353849704927347,
|
| 1103 |
+
"learning_rate": 8.252525252525253e-05,
|
| 1104 |
+
"loss": 2.0225625,
|
| 1105 |
+
"step": 183000
|
| 1106 |
+
},
|
| 1107 |
+
{
|
| 1108 |
+
"epoch": 0.13612477907466222,
|
| 1109 |
+
"learning_rate": 8.242424242424243e-05,
|
| 1110 |
+
"loss": 2.02978125,
|
| 1111 |
+
"step": 184000
|
| 1112 |
+
},
|
| 1113 |
+
{
|
| 1114 |
+
"epoch": 0.13686458765658974,
|
| 1115 |
+
"learning_rate": 8.232323232323233e-05,
|
| 1116 |
+
"loss": 2.09690625,
|
| 1117 |
+
"step": 185000
|
| 1118 |
+
},
|
| 1119 |
+
{
|
| 1120 |
+
"epoch": 0.13760439623851725,
|
| 1121 |
+
"learning_rate": 8.222222222222222e-05,
|
| 1122 |
+
"loss": 2.14709375,
|
| 1123 |
+
"step": 186000
|
| 1124 |
+
},
|
| 1125 |
+
{
|
| 1126 |
+
"epoch": 0.13834420482044477,
|
| 1127 |
+
"learning_rate": 8.212121212121212e-05,
|
| 1128 |
+
"loss": 2.05809375,
|
| 1129 |
+
"step": 187000
|
| 1130 |
+
},
|
| 1131 |
+
{
|
| 1132 |
+
"epoch": 0.13908401340237228,
|
| 1133 |
+
"learning_rate": 8.202020202020202e-05,
|
| 1134 |
+
"loss": 2.24625,
|
| 1135 |
+
"step": 188000
|
| 1136 |
+
},
|
| 1137 |
+
{
|
| 1138 |
+
"epoch": 0.1398238219842998,
|
| 1139 |
+
"learning_rate": 8.191919191919192e-05,
|
| 1140 |
+
"loss": 2.2840625,
|
| 1141 |
+
"step": 189000
|
| 1142 |
+
},
|
| 1143 |
+
{
|
| 1144 |
+
"epoch": 0.14056363056622728,
|
| 1145 |
+
"learning_rate": 8.181818181818183e-05,
|
| 1146 |
+
"loss": 2.23384375,
|
| 1147 |
+
"step": 190000
|
| 1148 |
+
},
|
| 1149 |
+
{
|
| 1150 |
+
"epoch": 0.1413034391481548,
|
| 1151 |
+
"learning_rate": 8.171717171717172e-05,
|
| 1152 |
+
"loss": 2.27740625,
|
| 1153 |
+
"step": 191000
|
| 1154 |
+
},
|
| 1155 |
+
{
|
| 1156 |
+
"epoch": 0.1420432477300823,
|
| 1157 |
+
"learning_rate": 8.161616161616163e-05,
|
| 1158 |
+
"loss": 2.3976875,
|
| 1159 |
+
"step": 192000
|
| 1160 |
+
},
|
| 1161 |
+
{
|
| 1162 |
+
"epoch": 0.14278305631200983,
|
| 1163 |
+
"learning_rate": 8.151515151515152e-05,
|
| 1164 |
+
"loss": 2.22225,
|
| 1165 |
+
"step": 193000
|
| 1166 |
+
},
|
| 1167 |
+
{
|
| 1168 |
+
"epoch": 0.14352286489393734,
|
| 1169 |
+
"learning_rate": 8.141414141414141e-05,
|
| 1170 |
+
"loss": 2.28578125,
|
| 1171 |
+
"step": 194000
|
| 1172 |
+
},
|
| 1173 |
+
{
|
| 1174 |
+
"epoch": 0.14426267347586486,
|
| 1175 |
+
"learning_rate": 8.131313131313132e-05,
|
| 1176 |
+
"loss": 2.22359375,
|
| 1177 |
+
"step": 195000
|
| 1178 |
+
},
|
| 1179 |
+
{
|
| 1180 |
+
"epoch": 0.14500248205779237,
|
| 1181 |
+
"learning_rate": 8.121212121212121e-05,
|
| 1182 |
+
"loss": 2.16703125,
|
| 1183 |
+
"step": 196000
|
| 1184 |
+
},
|
| 1185 |
+
{
|
| 1186 |
+
"epoch": 0.14574229063971988,
|
| 1187 |
+
"learning_rate": 8.111111111111112e-05,
|
| 1188 |
+
"loss": 2.26603125,
|
| 1189 |
+
"step": 197000
|
| 1190 |
+
},
|
| 1191 |
+
{
|
| 1192 |
+
"epoch": 0.1464820992216474,
|
| 1193 |
+
"learning_rate": 8.101010101010101e-05,
|
| 1194 |
+
"loss": 2.159125,
|
| 1195 |
+
"step": 198000
|
| 1196 |
+
},
|
| 1197 |
+
{
|
| 1198 |
+
"epoch": 0.14722190780357491,
|
| 1199 |
+
"learning_rate": 8.090909090909092e-05,
|
| 1200 |
+
"loss": 2.02203125,
|
| 1201 |
+
"step": 199000
|
| 1202 |
+
},
|
| 1203 |
+
{
|
| 1204 |
+
"epoch": 0.14796171638550243,
|
| 1205 |
+
"learning_rate": 8.080808080808081e-05,
|
| 1206 |
+
"loss": 2.01090625,
|
| 1207 |
+
"step": 200000
|
| 1208 |
+
},
|
| 1209 |
+
{
|
| 1210 |
+
"epoch": 0.14870152496742992,
|
| 1211 |
+
"learning_rate": 8.07070707070707e-05,
|
| 1212 |
+
"loss": 2.0288125,
|
| 1213 |
+
"step": 201000
|
| 1214 |
+
},
|
| 1215 |
+
{
|
| 1216 |
+
"epoch": 0.14944133354935743,
|
| 1217 |
+
"learning_rate": 8.060606060606061e-05,
|
| 1218 |
+
"loss": 2.00403125,
|
| 1219 |
+
"step": 202000
|
| 1220 |
+
},
|
| 1221 |
+
{
|
| 1222 |
+
"epoch": 0.15018114213128494,
|
| 1223 |
+
"learning_rate": 8.05050505050505e-05,
|
| 1224 |
+
"loss": 1.99765625,
|
| 1225 |
+
"step": 203000
|
| 1226 |
+
},
|
| 1227 |
+
{
|
| 1228 |
+
"epoch": 0.15092095071321246,
|
| 1229 |
+
"learning_rate": 8.040404040404041e-05,
|
| 1230 |
+
"loss": 1.98784375,
|
| 1231 |
+
"step": 204000
|
| 1232 |
+
}
|
| 1233 |
+
],
|
| 1234 |
+
"max_steps": 1000000,
|
| 1235 |
+
"num_train_epochs": 1,
|
| 1236 |
+
"total_flos": 369792323308800000
|
| 1237 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:12d8e51be8872528fe91391a881d853bc49eac3e20c3208c9e2dc981e4968ad2
|
| 3 |
+
size 1839
|
vocab.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|