Achmat Sodikkun commited on
Commit ·
1e78c18
1
Parent(s): 28b2b10
feat: add model finetuning pantun
Browse filesThis view is limited to 50 files because it contains too many changes.
See raw diff
- added_tokens.json +5 -0
- checkpoint-1000/config.json +40 -0
- checkpoint-1000/generation_config.json +6 -0
- checkpoint-1000/model.safetensors +3 -0
- checkpoint-1000/optimizer.pt +3 -0
- checkpoint-1000/rng_state.pth +3 -0
- checkpoint-1000/scheduler.pt +3 -0
- checkpoint-1000/trainer_state.json +1015 -0
- checkpoint-1000/training_args.bin +3 -0
- checkpoint-1500/config.json +40 -0
- checkpoint-1500/generation_config.json +6 -0
- checkpoint-1500/model.safetensors +3 -0
- checkpoint-1500/optimizer.pt +3 -0
- checkpoint-1500/rng_state.pth +3 -0
- checkpoint-1500/scheduler.pt +3 -0
- checkpoint-1500/trainer_state.json +1519 -0
- checkpoint-1500/training_args.bin +3 -0
- checkpoint-2000/config.json +40 -0
- checkpoint-2000/generation_config.json +6 -0
- checkpoint-2000/model.safetensors +3 -0
- checkpoint-2000/optimizer.pt +3 -0
- checkpoint-2000/rng_state.pth +3 -0
- checkpoint-2000/scheduler.pt +3 -0
- checkpoint-2000/trainer_state.json +2017 -0
- checkpoint-2000/training_args.bin +3 -0
- checkpoint-2500/config.json +40 -0
- checkpoint-2500/generation_config.json +6 -0
- checkpoint-2500/model.safetensors +3 -0
- checkpoint-2500/optimizer.pt +3 -0
- checkpoint-2500/rng_state.pth +3 -0
- checkpoint-2500/scheduler.pt +3 -0
- checkpoint-2500/trainer_state.json +2515 -0
- checkpoint-2500/training_args.bin +3 -0
- checkpoint-3000/config.json +40 -0
- checkpoint-3000/generation_config.json +6 -0
- checkpoint-3000/model.safetensors +3 -0
- checkpoint-3000/optimizer.pt +3 -0
- checkpoint-3000/rng_state.pth +3 -0
- checkpoint-3000/scheduler.pt +3 -0
- checkpoint-3000/trainer_state.json +3019 -0
- checkpoint-3000/training_args.bin +3 -0
- checkpoint-500/config.json +40 -0
- checkpoint-500/generation_config.json +6 -0
- checkpoint-500/model.safetensors +3 -0
- checkpoint-500/optimizer.pt +3 -0
- checkpoint-500/rng_state.pth +3 -0
- checkpoint-500/scheduler.pt +3 -0
- checkpoint-500/trainer_state.json +517 -0
- checkpoint-500/training_args.bin +3 -0
- config.json +40 -0
added_tokens.json
ADDED
|
@@ -0,0 +1,5 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<BOS>": 50257,
|
| 3 |
+
"<EOS>": 50258,
|
| 4 |
+
"<PAD>": 50259
|
| 5 |
+
}
|
checkpoint-1000/config.json
ADDED
|
@@ -0,0 +1,40 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "flax-community/gpt2-small-indonesian",
|
| 3 |
+
"activation_function": "gelu_new",
|
| 4 |
+
"architectures": [
|
| 5 |
+
"GPT2LMHeadModel"
|
| 6 |
+
],
|
| 7 |
+
"attn_pdrop": 0.0,
|
| 8 |
+
"bos_token_id": 50256,
|
| 9 |
+
"embd_pdrop": 0.0,
|
| 10 |
+
"eos_token_id": 50256,
|
| 11 |
+
"gradient_checkpointing": false,
|
| 12 |
+
"initializer_range": 0.02,
|
| 13 |
+
"layer_norm_epsilon": 1e-05,
|
| 14 |
+
"model_type": "gpt2",
|
| 15 |
+
"n_ctx": 1024,
|
| 16 |
+
"n_embd": 768,
|
| 17 |
+
"n_head": 12,
|
| 18 |
+
"n_inner": null,
|
| 19 |
+
"n_layer": 12,
|
| 20 |
+
"n_positions": 1024,
|
| 21 |
+
"reorder_and_upcast_attn": false,
|
| 22 |
+
"resid_pdrop": 0.0,
|
| 23 |
+
"scale_attn_by_inverse_layer_idx": false,
|
| 24 |
+
"scale_attn_weights": true,
|
| 25 |
+
"summary_activation": null,
|
| 26 |
+
"summary_first_dropout": 0.1,
|
| 27 |
+
"summary_proj_to_labels": true,
|
| 28 |
+
"summary_type": "cls_index",
|
| 29 |
+
"summary_use_proj": true,
|
| 30 |
+
"task_specific_params": {
|
| 31 |
+
"text-generation": {
|
| 32 |
+
"do_sample": true,
|
| 33 |
+
"max_length": 50
|
| 34 |
+
}
|
| 35 |
+
},
|
| 36 |
+
"torch_dtype": "float32",
|
| 37 |
+
"transformers_version": "4.35.2",
|
| 38 |
+
"use_cache": true,
|
| 39 |
+
"vocab_size": 50260
|
| 40 |
+
}
|
checkpoint-1000/generation_config.json
ADDED
|
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"bos_token_id": 50256,
|
| 4 |
+
"eos_token_id": 50256,
|
| 5 |
+
"transformers_version": "4.35.2"
|
| 6 |
+
}
|
checkpoint-1000/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:05c57f729236dc7a335b54a5e5a115d9145147253045fa3f825d267d43fdfbcb
|
| 3 |
+
size 497783424
|
checkpoint-1000/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9342ac2c0f582cd8b74e7bc42c505a2a57c2c5c43779a879541b5861c3caa867
|
| 3 |
+
size 995660730
|
checkpoint-1000/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9899ccda7f0d8d9511991180b93aab508ce6e8489de708c88ad1188e7e1d90d6
|
| 3 |
+
size 14244
|
checkpoint-1000/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a8f8a566a9f9a45a4b99ab20b6e009c8082631d17c87a8869b9dbd71f79967b6
|
| 3 |
+
size 1064
|
checkpoint-1000/trainer_state.json
ADDED
|
@@ -0,0 +1,1015 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 3.2051282051282053,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 1000,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.02,
|
| 13 |
+
"learning_rate": 4.9903846153846154e-05,
|
| 14 |
+
"loss": 5.8218,
|
| 15 |
+
"step": 6
|
| 16 |
+
},
|
| 17 |
+
{
|
| 18 |
+
"epoch": 0.04,
|
| 19 |
+
"learning_rate": 4.980769230769231e-05,
|
| 20 |
+
"loss": 5.2499,
|
| 21 |
+
"step": 12
|
| 22 |
+
},
|
| 23 |
+
{
|
| 24 |
+
"epoch": 0.06,
|
| 25 |
+
"learning_rate": 4.9711538461538465e-05,
|
| 26 |
+
"loss": 5.1236,
|
| 27 |
+
"step": 18
|
| 28 |
+
},
|
| 29 |
+
{
|
| 30 |
+
"epoch": 0.08,
|
| 31 |
+
"learning_rate": 4.961538461538462e-05,
|
| 32 |
+
"loss": 4.8363,
|
| 33 |
+
"step": 24
|
| 34 |
+
},
|
| 35 |
+
{
|
| 36 |
+
"epoch": 0.1,
|
| 37 |
+
"learning_rate": 4.9519230769230776e-05,
|
| 38 |
+
"loss": 4.5805,
|
| 39 |
+
"step": 30
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.12,
|
| 43 |
+
"learning_rate": 4.942307692307693e-05,
|
| 44 |
+
"loss": 4.5276,
|
| 45 |
+
"step": 36
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.13,
|
| 49 |
+
"learning_rate": 4.932692307692308e-05,
|
| 50 |
+
"loss": 4.3871,
|
| 51 |
+
"step": 42
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.15,
|
| 55 |
+
"learning_rate": 4.923076923076924e-05,
|
| 56 |
+
"loss": 4.2706,
|
| 57 |
+
"step": 48
|
| 58 |
+
},
|
| 59 |
+
{
|
| 60 |
+
"epoch": 0.17,
|
| 61 |
+
"learning_rate": 4.9134615384615384e-05,
|
| 62 |
+
"loss": 4.0906,
|
| 63 |
+
"step": 54
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 0.19,
|
| 67 |
+
"learning_rate": 4.9038461538461536e-05,
|
| 68 |
+
"loss": 4.1704,
|
| 69 |
+
"step": 60
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 0.21,
|
| 73 |
+
"learning_rate": 4.8942307692307695e-05,
|
| 74 |
+
"loss": 3.9014,
|
| 75 |
+
"step": 66
|
| 76 |
+
},
|
| 77 |
+
{
|
| 78 |
+
"epoch": 0.23,
|
| 79 |
+
"learning_rate": 4.884615384615385e-05,
|
| 80 |
+
"loss": 4.0338,
|
| 81 |
+
"step": 72
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"epoch": 0.25,
|
| 85 |
+
"learning_rate": 4.875e-05,
|
| 86 |
+
"loss": 4.011,
|
| 87 |
+
"step": 78
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"epoch": 0.27,
|
| 91 |
+
"learning_rate": 4.865384615384616e-05,
|
| 92 |
+
"loss": 3.8017,
|
| 93 |
+
"step": 84
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.29,
|
| 97 |
+
"learning_rate": 4.855769230769231e-05,
|
| 98 |
+
"loss": 3.819,
|
| 99 |
+
"step": 90
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"epoch": 0.31,
|
| 103 |
+
"learning_rate": 4.846153846153846e-05,
|
| 104 |
+
"loss": 3.8157,
|
| 105 |
+
"step": 96
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 0.33,
|
| 109 |
+
"learning_rate": 4.836538461538462e-05,
|
| 110 |
+
"loss": 3.7675,
|
| 111 |
+
"step": 102
|
| 112 |
+
},
|
| 113 |
+
{
|
| 114 |
+
"epoch": 0.35,
|
| 115 |
+
"learning_rate": 4.826923076923077e-05,
|
| 116 |
+
"loss": 3.756,
|
| 117 |
+
"step": 108
|
| 118 |
+
},
|
| 119 |
+
{
|
| 120 |
+
"epoch": 0.37,
|
| 121 |
+
"learning_rate": 4.8173076923076925e-05,
|
| 122 |
+
"loss": 3.7439,
|
| 123 |
+
"step": 114
|
| 124 |
+
},
|
| 125 |
+
{
|
| 126 |
+
"epoch": 0.38,
|
| 127 |
+
"learning_rate": 4.8076923076923084e-05,
|
| 128 |
+
"loss": 3.761,
|
| 129 |
+
"step": 120
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.4,
|
| 133 |
+
"learning_rate": 4.7980769230769236e-05,
|
| 134 |
+
"loss": 3.7318,
|
| 135 |
+
"step": 126
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.42,
|
| 139 |
+
"learning_rate": 4.788461538461539e-05,
|
| 140 |
+
"loss": 3.745,
|
| 141 |
+
"step": 132
|
| 142 |
+
},
|
| 143 |
+
{
|
| 144 |
+
"epoch": 0.44,
|
| 145 |
+
"learning_rate": 4.778846153846154e-05,
|
| 146 |
+
"loss": 3.784,
|
| 147 |
+
"step": 138
|
| 148 |
+
},
|
| 149 |
+
{
|
| 150 |
+
"epoch": 0.46,
|
| 151 |
+
"learning_rate": 4.76923076923077e-05,
|
| 152 |
+
"loss": 3.6114,
|
| 153 |
+
"step": 144
|
| 154 |
+
},
|
| 155 |
+
{
|
| 156 |
+
"epoch": 0.48,
|
| 157 |
+
"learning_rate": 4.7596153846153844e-05,
|
| 158 |
+
"loss": 3.637,
|
| 159 |
+
"step": 150
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 0.5,
|
| 163 |
+
"learning_rate": 4.75e-05,
|
| 164 |
+
"loss": 3.5881,
|
| 165 |
+
"step": 156
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 0.52,
|
| 169 |
+
"learning_rate": 4.7403846153846155e-05,
|
| 170 |
+
"loss": 3.5336,
|
| 171 |
+
"step": 162
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 0.54,
|
| 175 |
+
"learning_rate": 4.730769230769231e-05,
|
| 176 |
+
"loss": 3.5313,
|
| 177 |
+
"step": 168
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.56,
|
| 181 |
+
"learning_rate": 4.7211538461538465e-05,
|
| 182 |
+
"loss": 3.6357,
|
| 183 |
+
"step": 174
|
| 184 |
+
},
|
| 185 |
+
{
|
| 186 |
+
"epoch": 0.58,
|
| 187 |
+
"learning_rate": 4.711538461538462e-05,
|
| 188 |
+
"loss": 3.5818,
|
| 189 |
+
"step": 180
|
| 190 |
+
},
|
| 191 |
+
{
|
| 192 |
+
"epoch": 0.6,
|
| 193 |
+
"learning_rate": 4.701923076923077e-05,
|
| 194 |
+
"loss": 3.5017,
|
| 195 |
+
"step": 186
|
| 196 |
+
},
|
| 197 |
+
{
|
| 198 |
+
"epoch": 0.62,
|
| 199 |
+
"learning_rate": 4.692307692307693e-05,
|
| 200 |
+
"loss": 3.5815,
|
| 201 |
+
"step": 192
|
| 202 |
+
},
|
| 203 |
+
{
|
| 204 |
+
"epoch": 0.63,
|
| 205 |
+
"learning_rate": 4.682692307692308e-05,
|
| 206 |
+
"loss": 3.611,
|
| 207 |
+
"step": 198
|
| 208 |
+
},
|
| 209 |
+
{
|
| 210 |
+
"epoch": 0.65,
|
| 211 |
+
"learning_rate": 4.673076923076923e-05,
|
| 212 |
+
"loss": 3.6096,
|
| 213 |
+
"step": 204
|
| 214 |
+
},
|
| 215 |
+
{
|
| 216 |
+
"epoch": 0.67,
|
| 217 |
+
"learning_rate": 4.6634615384615384e-05,
|
| 218 |
+
"loss": 3.5436,
|
| 219 |
+
"step": 210
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.69,
|
| 223 |
+
"learning_rate": 4.653846153846154e-05,
|
| 224 |
+
"loss": 3.5258,
|
| 225 |
+
"step": 216
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 0.71,
|
| 229 |
+
"learning_rate": 4.6442307692307695e-05,
|
| 230 |
+
"loss": 3.5277,
|
| 231 |
+
"step": 222
|
| 232 |
+
},
|
| 233 |
+
{
|
| 234 |
+
"epoch": 0.73,
|
| 235 |
+
"learning_rate": 4.634615384615385e-05,
|
| 236 |
+
"loss": 3.505,
|
| 237 |
+
"step": 228
|
| 238 |
+
},
|
| 239 |
+
{
|
| 240 |
+
"epoch": 0.75,
|
| 241 |
+
"learning_rate": 4.6250000000000006e-05,
|
| 242 |
+
"loss": 3.4665,
|
| 243 |
+
"step": 234
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"epoch": 0.77,
|
| 247 |
+
"learning_rate": 4.615384615384616e-05,
|
| 248 |
+
"loss": 3.4245,
|
| 249 |
+
"step": 240
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 0.79,
|
| 253 |
+
"learning_rate": 4.605769230769231e-05,
|
| 254 |
+
"loss": 3.4547,
|
| 255 |
+
"step": 246
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 0.81,
|
| 259 |
+
"learning_rate": 4.596153846153846e-05,
|
| 260 |
+
"loss": 3.4382,
|
| 261 |
+
"step": 252
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.83,
|
| 265 |
+
"learning_rate": 4.5865384615384614e-05,
|
| 266 |
+
"loss": 3.482,
|
| 267 |
+
"step": 258
|
| 268 |
+
},
|
| 269 |
+
{
|
| 270 |
+
"epoch": 0.85,
|
| 271 |
+
"learning_rate": 4.576923076923077e-05,
|
| 272 |
+
"loss": 3.4795,
|
| 273 |
+
"step": 264
|
| 274 |
+
},
|
| 275 |
+
{
|
| 276 |
+
"epoch": 0.87,
|
| 277 |
+
"learning_rate": 4.5673076923076925e-05,
|
| 278 |
+
"loss": 3.5018,
|
| 279 |
+
"step": 270
|
| 280 |
+
},
|
| 281 |
+
{
|
| 282 |
+
"epoch": 0.88,
|
| 283 |
+
"learning_rate": 4.557692307692308e-05,
|
| 284 |
+
"loss": 3.448,
|
| 285 |
+
"step": 276
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 0.9,
|
| 289 |
+
"learning_rate": 4.548076923076923e-05,
|
| 290 |
+
"loss": 3.4206,
|
| 291 |
+
"step": 282
|
| 292 |
+
},
|
| 293 |
+
{
|
| 294 |
+
"epoch": 0.92,
|
| 295 |
+
"learning_rate": 4.538461538461539e-05,
|
| 296 |
+
"loss": 3.528,
|
| 297 |
+
"step": 288
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 0.94,
|
| 301 |
+
"learning_rate": 4.528846153846154e-05,
|
| 302 |
+
"loss": 3.475,
|
| 303 |
+
"step": 294
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.96,
|
| 307 |
+
"learning_rate": 4.519230769230769e-05,
|
| 308 |
+
"loss": 3.4142,
|
| 309 |
+
"step": 300
|
| 310 |
+
},
|
| 311 |
+
{
|
| 312 |
+
"epoch": 0.98,
|
| 313 |
+
"learning_rate": 4.509615384615385e-05,
|
| 314 |
+
"loss": 3.4156,
|
| 315 |
+
"step": 306
|
| 316 |
+
},
|
| 317 |
+
{
|
| 318 |
+
"epoch": 1.0,
|
| 319 |
+
"learning_rate": 4.5e-05,
|
| 320 |
+
"loss": 3.3274,
|
| 321 |
+
"step": 312
|
| 322 |
+
},
|
| 323 |
+
{
|
| 324 |
+
"epoch": 1.02,
|
| 325 |
+
"learning_rate": 4.4903846153846155e-05,
|
| 326 |
+
"loss": 3.3594,
|
| 327 |
+
"step": 318
|
| 328 |
+
},
|
| 329 |
+
{
|
| 330 |
+
"epoch": 1.04,
|
| 331 |
+
"learning_rate": 4.4807692307692314e-05,
|
| 332 |
+
"loss": 3.2517,
|
| 333 |
+
"step": 324
|
| 334 |
+
},
|
| 335 |
+
{
|
| 336 |
+
"epoch": 1.06,
|
| 337 |
+
"learning_rate": 4.4711538461538466e-05,
|
| 338 |
+
"loss": 3.2934,
|
| 339 |
+
"step": 330
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 1.08,
|
| 343 |
+
"learning_rate": 4.461538461538462e-05,
|
| 344 |
+
"loss": 3.2986,
|
| 345 |
+
"step": 336
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 1.1,
|
| 349 |
+
"learning_rate": 4.451923076923077e-05,
|
| 350 |
+
"loss": 3.2655,
|
| 351 |
+
"step": 342
|
| 352 |
+
},
|
| 353 |
+
{
|
| 354 |
+
"epoch": 1.12,
|
| 355 |
+
"learning_rate": 4.442307692307692e-05,
|
| 356 |
+
"loss": 3.4058,
|
| 357 |
+
"step": 348
|
| 358 |
+
},
|
| 359 |
+
{
|
| 360 |
+
"epoch": 1.13,
|
| 361 |
+
"learning_rate": 4.4326923076923074e-05,
|
| 362 |
+
"loss": 3.265,
|
| 363 |
+
"step": 354
|
| 364 |
+
},
|
| 365 |
+
{
|
| 366 |
+
"epoch": 1.15,
|
| 367 |
+
"learning_rate": 4.423076923076923e-05,
|
| 368 |
+
"loss": 3.3208,
|
| 369 |
+
"step": 360
|
| 370 |
+
},
|
| 371 |
+
{
|
| 372 |
+
"epoch": 1.17,
|
| 373 |
+
"learning_rate": 4.4134615384615385e-05,
|
| 374 |
+
"loss": 3.238,
|
| 375 |
+
"step": 366
|
| 376 |
+
},
|
| 377 |
+
{
|
| 378 |
+
"epoch": 1.19,
|
| 379 |
+
"learning_rate": 4.403846153846154e-05,
|
| 380 |
+
"loss": 3.2153,
|
| 381 |
+
"step": 372
|
| 382 |
+
},
|
| 383 |
+
{
|
| 384 |
+
"epoch": 1.21,
|
| 385 |
+
"learning_rate": 4.3942307692307695e-05,
|
| 386 |
+
"loss": 3.264,
|
| 387 |
+
"step": 378
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 1.23,
|
| 391 |
+
"learning_rate": 4.384615384615385e-05,
|
| 392 |
+
"loss": 3.2911,
|
| 393 |
+
"step": 384
|
| 394 |
+
},
|
| 395 |
+
{
|
| 396 |
+
"epoch": 1.25,
|
| 397 |
+
"learning_rate": 4.375e-05,
|
| 398 |
+
"loss": 3.3027,
|
| 399 |
+
"step": 390
|
| 400 |
+
},
|
| 401 |
+
{
|
| 402 |
+
"epoch": 1.27,
|
| 403 |
+
"learning_rate": 4.365384615384616e-05,
|
| 404 |
+
"loss": 3.2589,
|
| 405 |
+
"step": 396
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 1.29,
|
| 409 |
+
"learning_rate": 4.355769230769231e-05,
|
| 410 |
+
"loss": 3.3683,
|
| 411 |
+
"step": 402
|
| 412 |
+
},
|
| 413 |
+
{
|
| 414 |
+
"epoch": 1.31,
|
| 415 |
+
"learning_rate": 4.346153846153846e-05,
|
| 416 |
+
"loss": 3.2849,
|
| 417 |
+
"step": 408
|
| 418 |
+
},
|
| 419 |
+
{
|
| 420 |
+
"epoch": 1.33,
|
| 421 |
+
"learning_rate": 4.336538461538462e-05,
|
| 422 |
+
"loss": 3.2397,
|
| 423 |
+
"step": 414
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 1.35,
|
| 427 |
+
"learning_rate": 4.326923076923077e-05,
|
| 428 |
+
"loss": 3.2128,
|
| 429 |
+
"step": 420
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 1.37,
|
| 433 |
+
"learning_rate": 4.3173076923076925e-05,
|
| 434 |
+
"loss": 3.1944,
|
| 435 |
+
"step": 426
|
| 436 |
+
},
|
| 437 |
+
{
|
| 438 |
+
"epoch": 1.38,
|
| 439 |
+
"learning_rate": 4.3076923076923084e-05,
|
| 440 |
+
"loss": 3.1837,
|
| 441 |
+
"step": 432
|
| 442 |
+
},
|
| 443 |
+
{
|
| 444 |
+
"epoch": 1.4,
|
| 445 |
+
"learning_rate": 4.2980769230769236e-05,
|
| 446 |
+
"loss": 3.1793,
|
| 447 |
+
"step": 438
|
| 448 |
+
},
|
| 449 |
+
{
|
| 450 |
+
"epoch": 1.42,
|
| 451 |
+
"learning_rate": 4.288461538461538e-05,
|
| 452 |
+
"loss": 3.1447,
|
| 453 |
+
"step": 444
|
| 454 |
+
},
|
| 455 |
+
{
|
| 456 |
+
"epoch": 1.44,
|
| 457 |
+
"learning_rate": 4.278846153846154e-05,
|
| 458 |
+
"loss": 3.1028,
|
| 459 |
+
"step": 450
|
| 460 |
+
},
|
| 461 |
+
{
|
| 462 |
+
"epoch": 1.46,
|
| 463 |
+
"learning_rate": 4.269230769230769e-05,
|
| 464 |
+
"loss": 3.2471,
|
| 465 |
+
"step": 456
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 1.48,
|
| 469 |
+
"learning_rate": 4.2596153846153844e-05,
|
| 470 |
+
"loss": 3.1855,
|
| 471 |
+
"step": 462
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 1.5,
|
| 475 |
+
"learning_rate": 4.25e-05,
|
| 476 |
+
"loss": 3.1817,
|
| 477 |
+
"step": 468
|
| 478 |
+
},
|
| 479 |
+
{
|
| 480 |
+
"epoch": 1.52,
|
| 481 |
+
"learning_rate": 4.2403846153846155e-05,
|
| 482 |
+
"loss": 3.214,
|
| 483 |
+
"step": 474
|
| 484 |
+
},
|
| 485 |
+
{
|
| 486 |
+
"epoch": 1.54,
|
| 487 |
+
"learning_rate": 4.230769230769231e-05,
|
| 488 |
+
"loss": 2.997,
|
| 489 |
+
"step": 480
|
| 490 |
+
},
|
| 491 |
+
{
|
| 492 |
+
"epoch": 1.56,
|
| 493 |
+
"learning_rate": 4.2211538461538466e-05,
|
| 494 |
+
"loss": 3.2059,
|
| 495 |
+
"step": 486
|
| 496 |
+
},
|
| 497 |
+
{
|
| 498 |
+
"epoch": 1.58,
|
| 499 |
+
"learning_rate": 4.211538461538462e-05,
|
| 500 |
+
"loss": 3.1517,
|
| 501 |
+
"step": 492
|
| 502 |
+
},
|
| 503 |
+
{
|
| 504 |
+
"epoch": 1.6,
|
| 505 |
+
"learning_rate": 4.201923076923077e-05,
|
| 506 |
+
"loss": 3.2236,
|
| 507 |
+
"step": 498
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 1.62,
|
| 511 |
+
"learning_rate": 4.192307692307693e-05,
|
| 512 |
+
"loss": 3.1939,
|
| 513 |
+
"step": 504
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 1.63,
|
| 517 |
+
"learning_rate": 4.182692307692308e-05,
|
| 518 |
+
"loss": 3.1143,
|
| 519 |
+
"step": 510
|
| 520 |
+
},
|
| 521 |
+
{
|
| 522 |
+
"epoch": 1.65,
|
| 523 |
+
"learning_rate": 4.173076923076923e-05,
|
| 524 |
+
"loss": 3.2261,
|
| 525 |
+
"step": 516
|
| 526 |
+
},
|
| 527 |
+
{
|
| 528 |
+
"epoch": 1.67,
|
| 529 |
+
"learning_rate": 4.163461538461539e-05,
|
| 530 |
+
"loss": 3.2095,
|
| 531 |
+
"step": 522
|
| 532 |
+
},
|
| 533 |
+
{
|
| 534 |
+
"epoch": 1.69,
|
| 535 |
+
"learning_rate": 4.1538461538461544e-05,
|
| 536 |
+
"loss": 3.2498,
|
| 537 |
+
"step": 528
|
| 538 |
+
},
|
| 539 |
+
{
|
| 540 |
+
"epoch": 1.71,
|
| 541 |
+
"learning_rate": 4.1442307692307696e-05,
|
| 542 |
+
"loss": 3.1493,
|
| 543 |
+
"step": 534
|
| 544 |
+
},
|
| 545 |
+
{
|
| 546 |
+
"epoch": 1.73,
|
| 547 |
+
"learning_rate": 4.134615384615385e-05,
|
| 548 |
+
"loss": 3.0919,
|
| 549 |
+
"step": 540
|
| 550 |
+
},
|
| 551 |
+
{
|
| 552 |
+
"epoch": 1.75,
|
| 553 |
+
"learning_rate": 4.125e-05,
|
| 554 |
+
"loss": 3.2617,
|
| 555 |
+
"step": 546
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 1.77,
|
| 559 |
+
"learning_rate": 4.115384615384615e-05,
|
| 560 |
+
"loss": 3.1447,
|
| 561 |
+
"step": 552
|
| 562 |
+
},
|
| 563 |
+
{
|
| 564 |
+
"epoch": 1.79,
|
| 565 |
+
"learning_rate": 4.105769230769231e-05,
|
| 566 |
+
"loss": 3.2025,
|
| 567 |
+
"step": 558
|
| 568 |
+
},
|
| 569 |
+
{
|
| 570 |
+
"epoch": 1.81,
|
| 571 |
+
"learning_rate": 4.096153846153846e-05,
|
| 572 |
+
"loss": 3.1829,
|
| 573 |
+
"step": 564
|
| 574 |
+
},
|
| 575 |
+
{
|
| 576 |
+
"epoch": 1.83,
|
| 577 |
+
"learning_rate": 4.0865384615384615e-05,
|
| 578 |
+
"loss": 3.2216,
|
| 579 |
+
"step": 570
|
| 580 |
+
},
|
| 581 |
+
{
|
| 582 |
+
"epoch": 1.85,
|
| 583 |
+
"learning_rate": 4.0769230769230773e-05,
|
| 584 |
+
"loss": 3.1391,
|
| 585 |
+
"step": 576
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 1.87,
|
| 589 |
+
"learning_rate": 4.0673076923076926e-05,
|
| 590 |
+
"loss": 3.1344,
|
| 591 |
+
"step": 582
|
| 592 |
+
},
|
| 593 |
+
{
|
| 594 |
+
"epoch": 1.88,
|
| 595 |
+
"learning_rate": 4.057692307692308e-05,
|
| 596 |
+
"loss": 3.1699,
|
| 597 |
+
"step": 588
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 1.9,
|
| 601 |
+
"learning_rate": 4.0480769230769236e-05,
|
| 602 |
+
"loss": 3.0138,
|
| 603 |
+
"step": 594
|
| 604 |
+
},
|
| 605 |
+
{
|
| 606 |
+
"epoch": 1.92,
|
| 607 |
+
"learning_rate": 4.038461538461539e-05,
|
| 608 |
+
"loss": 3.1157,
|
| 609 |
+
"step": 600
|
| 610 |
+
},
|
| 611 |
+
{
|
| 612 |
+
"epoch": 1.94,
|
| 613 |
+
"learning_rate": 4.028846153846154e-05,
|
| 614 |
+
"loss": 3.1173,
|
| 615 |
+
"step": 606
|
| 616 |
+
},
|
| 617 |
+
{
|
| 618 |
+
"epoch": 1.96,
|
| 619 |
+
"learning_rate": 4.019230769230769e-05,
|
| 620 |
+
"loss": 3.2019,
|
| 621 |
+
"step": 612
|
| 622 |
+
},
|
| 623 |
+
{
|
| 624 |
+
"epoch": 1.98,
|
| 625 |
+
"learning_rate": 4.009615384615385e-05,
|
| 626 |
+
"loss": 3.0801,
|
| 627 |
+
"step": 618
|
| 628 |
+
},
|
| 629 |
+
{
|
| 630 |
+
"epoch": 2.0,
|
| 631 |
+
"learning_rate": 4e-05,
|
| 632 |
+
"loss": 3.0888,
|
| 633 |
+
"step": 624
|
| 634 |
+
},
|
| 635 |
+
{
|
| 636 |
+
"epoch": 2.02,
|
| 637 |
+
"learning_rate": 3.9903846153846155e-05,
|
| 638 |
+
"loss": 3.0332,
|
| 639 |
+
"step": 630
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 2.04,
|
| 643 |
+
"learning_rate": 3.980769230769231e-05,
|
| 644 |
+
"loss": 2.9418,
|
| 645 |
+
"step": 636
|
| 646 |
+
},
|
| 647 |
+
{
|
| 648 |
+
"epoch": 2.06,
|
| 649 |
+
"learning_rate": 3.971153846153846e-05,
|
| 650 |
+
"loss": 2.933,
|
| 651 |
+
"step": 642
|
| 652 |
+
},
|
| 653 |
+
{
|
| 654 |
+
"epoch": 2.08,
|
| 655 |
+
"learning_rate": 3.961538461538462e-05,
|
| 656 |
+
"loss": 3.0467,
|
| 657 |
+
"step": 648
|
| 658 |
+
},
|
| 659 |
+
{
|
| 660 |
+
"epoch": 2.1,
|
| 661 |
+
"learning_rate": 3.951923076923077e-05,
|
| 662 |
+
"loss": 3.0991,
|
| 663 |
+
"step": 654
|
| 664 |
+
},
|
| 665 |
+
{
|
| 666 |
+
"epoch": 2.12,
|
| 667 |
+
"learning_rate": 3.942307692307692e-05,
|
| 668 |
+
"loss": 2.913,
|
| 669 |
+
"step": 660
|
| 670 |
+
},
|
| 671 |
+
{
|
| 672 |
+
"epoch": 2.13,
|
| 673 |
+
"learning_rate": 3.932692307692308e-05,
|
| 674 |
+
"loss": 3.0531,
|
| 675 |
+
"step": 666
|
| 676 |
+
},
|
| 677 |
+
{
|
| 678 |
+
"epoch": 2.15,
|
| 679 |
+
"learning_rate": 3.923076923076923e-05,
|
| 680 |
+
"loss": 2.9838,
|
| 681 |
+
"step": 672
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 2.17,
|
| 685 |
+
"learning_rate": 3.9134615384615385e-05,
|
| 686 |
+
"loss": 2.9406,
|
| 687 |
+
"step": 678
|
| 688 |
+
},
|
| 689 |
+
{
|
| 690 |
+
"epoch": 2.19,
|
| 691 |
+
"learning_rate": 3.903846153846154e-05,
|
| 692 |
+
"loss": 3.046,
|
| 693 |
+
"step": 684
|
| 694 |
+
},
|
| 695 |
+
{
|
| 696 |
+
"epoch": 2.21,
|
| 697 |
+
"learning_rate": 3.8942307692307696e-05,
|
| 698 |
+
"loss": 3.006,
|
| 699 |
+
"step": 690
|
| 700 |
+
},
|
| 701 |
+
{
|
| 702 |
+
"epoch": 2.23,
|
| 703 |
+
"learning_rate": 3.884615384615385e-05,
|
| 704 |
+
"loss": 2.9774,
|
| 705 |
+
"step": 696
|
| 706 |
+
},
|
| 707 |
+
{
|
| 708 |
+
"epoch": 2.25,
|
| 709 |
+
"learning_rate": 3.875e-05,
|
| 710 |
+
"loss": 2.9937,
|
| 711 |
+
"step": 702
|
| 712 |
+
},
|
| 713 |
+
{
|
| 714 |
+
"epoch": 2.27,
|
| 715 |
+
"learning_rate": 3.865384615384616e-05,
|
| 716 |
+
"loss": 2.9737,
|
| 717 |
+
"step": 708
|
| 718 |
+
},
|
| 719 |
+
{
|
| 720 |
+
"epoch": 2.29,
|
| 721 |
+
"learning_rate": 3.855769230769231e-05,
|
| 722 |
+
"loss": 2.9772,
|
| 723 |
+
"step": 714
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 2.31,
|
| 727 |
+
"learning_rate": 3.846153846153846e-05,
|
| 728 |
+
"loss": 2.8405,
|
| 729 |
+
"step": 720
|
| 730 |
+
},
|
| 731 |
+
{
|
| 732 |
+
"epoch": 2.33,
|
| 733 |
+
"learning_rate": 3.836538461538462e-05,
|
| 734 |
+
"loss": 3.0856,
|
| 735 |
+
"step": 726
|
| 736 |
+
},
|
| 737 |
+
{
|
| 738 |
+
"epoch": 2.35,
|
| 739 |
+
"learning_rate": 3.826923076923077e-05,
|
| 740 |
+
"loss": 2.9947,
|
| 741 |
+
"step": 732
|
| 742 |
+
},
|
| 743 |
+
{
|
| 744 |
+
"epoch": 2.37,
|
| 745 |
+
"learning_rate": 3.8173076923076926e-05,
|
| 746 |
+
"loss": 2.9251,
|
| 747 |
+
"step": 738
|
| 748 |
+
},
|
| 749 |
+
{
|
| 750 |
+
"epoch": 2.38,
|
| 751 |
+
"learning_rate": 3.807692307692308e-05,
|
| 752 |
+
"loss": 2.9613,
|
| 753 |
+
"step": 744
|
| 754 |
+
},
|
| 755 |
+
{
|
| 756 |
+
"epoch": 2.4,
|
| 757 |
+
"learning_rate": 3.798076923076923e-05,
|
| 758 |
+
"loss": 2.9402,
|
| 759 |
+
"step": 750
|
| 760 |
+
},
|
| 761 |
+
{
|
| 762 |
+
"epoch": 2.42,
|
| 763 |
+
"learning_rate": 3.788461538461538e-05,
|
| 764 |
+
"loss": 2.8823,
|
| 765 |
+
"step": 756
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 2.44,
|
| 769 |
+
"learning_rate": 3.778846153846154e-05,
|
| 770 |
+
"loss": 2.8906,
|
| 771 |
+
"step": 762
|
| 772 |
+
},
|
| 773 |
+
{
|
| 774 |
+
"epoch": 2.46,
|
| 775 |
+
"learning_rate": 3.769230769230769e-05,
|
| 776 |
+
"loss": 2.9475,
|
| 777 |
+
"step": 768
|
| 778 |
+
},
|
| 779 |
+
{
|
| 780 |
+
"epoch": 2.48,
|
| 781 |
+
"learning_rate": 3.7596153846153845e-05,
|
| 782 |
+
"loss": 2.9999,
|
| 783 |
+
"step": 774
|
| 784 |
+
},
|
| 785 |
+
{
|
| 786 |
+
"epoch": 2.5,
|
| 787 |
+
"learning_rate": 3.7500000000000003e-05,
|
| 788 |
+
"loss": 2.9503,
|
| 789 |
+
"step": 780
|
| 790 |
+
},
|
| 791 |
+
{
|
| 792 |
+
"epoch": 2.52,
|
| 793 |
+
"learning_rate": 3.7403846153846156e-05,
|
| 794 |
+
"loss": 3.0112,
|
| 795 |
+
"step": 786
|
| 796 |
+
},
|
| 797 |
+
{
|
| 798 |
+
"epoch": 2.54,
|
| 799 |
+
"learning_rate": 3.730769230769231e-05,
|
| 800 |
+
"loss": 3.0096,
|
| 801 |
+
"step": 792
|
| 802 |
+
},
|
| 803 |
+
{
|
| 804 |
+
"epoch": 2.56,
|
| 805 |
+
"learning_rate": 3.7211538461538466e-05,
|
| 806 |
+
"loss": 2.9804,
|
| 807 |
+
"step": 798
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 2.58,
|
| 811 |
+
"learning_rate": 3.711538461538462e-05,
|
| 812 |
+
"loss": 2.9203,
|
| 813 |
+
"step": 804
|
| 814 |
+
},
|
| 815 |
+
{
|
| 816 |
+
"epoch": 2.6,
|
| 817 |
+
"learning_rate": 3.701923076923077e-05,
|
| 818 |
+
"loss": 2.939,
|
| 819 |
+
"step": 810
|
| 820 |
+
},
|
| 821 |
+
{
|
| 822 |
+
"epoch": 2.62,
|
| 823 |
+
"learning_rate": 3.692307692307693e-05,
|
| 824 |
+
"loss": 2.8455,
|
| 825 |
+
"step": 816
|
| 826 |
+
},
|
| 827 |
+
{
|
| 828 |
+
"epoch": 2.63,
|
| 829 |
+
"learning_rate": 3.682692307692308e-05,
|
| 830 |
+
"loss": 2.9651,
|
| 831 |
+
"step": 822
|
| 832 |
+
},
|
| 833 |
+
{
|
| 834 |
+
"epoch": 2.65,
|
| 835 |
+
"learning_rate": 3.673076923076923e-05,
|
| 836 |
+
"loss": 2.9528,
|
| 837 |
+
"step": 828
|
| 838 |
+
},
|
| 839 |
+
{
|
| 840 |
+
"epoch": 2.67,
|
| 841 |
+
"learning_rate": 3.6634615384615385e-05,
|
| 842 |
+
"loss": 2.8042,
|
| 843 |
+
"step": 834
|
| 844 |
+
},
|
| 845 |
+
{
|
| 846 |
+
"epoch": 2.69,
|
| 847 |
+
"learning_rate": 3.653846153846154e-05,
|
| 848 |
+
"loss": 2.8311,
|
| 849 |
+
"step": 840
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 2.71,
|
| 853 |
+
"learning_rate": 3.644230769230769e-05,
|
| 854 |
+
"loss": 2.8888,
|
| 855 |
+
"step": 846
|
| 856 |
+
},
|
| 857 |
+
{
|
| 858 |
+
"epoch": 2.73,
|
| 859 |
+
"learning_rate": 3.634615384615385e-05,
|
| 860 |
+
"loss": 2.9151,
|
| 861 |
+
"step": 852
|
| 862 |
+
},
|
| 863 |
+
{
|
| 864 |
+
"epoch": 2.75,
|
| 865 |
+
"learning_rate": 3.625e-05,
|
| 866 |
+
"loss": 2.9463,
|
| 867 |
+
"step": 858
|
| 868 |
+
},
|
| 869 |
+
{
|
| 870 |
+
"epoch": 2.77,
|
| 871 |
+
"learning_rate": 3.615384615384615e-05,
|
| 872 |
+
"loss": 2.957,
|
| 873 |
+
"step": 864
|
| 874 |
+
},
|
| 875 |
+
{
|
| 876 |
+
"epoch": 2.79,
|
| 877 |
+
"learning_rate": 3.605769230769231e-05,
|
| 878 |
+
"loss": 2.9473,
|
| 879 |
+
"step": 870
|
| 880 |
+
},
|
| 881 |
+
{
|
| 882 |
+
"epoch": 2.81,
|
| 883 |
+
"learning_rate": 3.596153846153846e-05,
|
| 884 |
+
"loss": 2.9994,
|
| 885 |
+
"step": 876
|
| 886 |
+
},
|
| 887 |
+
{
|
| 888 |
+
"epoch": 2.83,
|
| 889 |
+
"learning_rate": 3.5865384615384615e-05,
|
| 890 |
+
"loss": 3.0486,
|
| 891 |
+
"step": 882
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 2.85,
|
| 895 |
+
"learning_rate": 3.5769230769230774e-05,
|
| 896 |
+
"loss": 2.9487,
|
| 897 |
+
"step": 888
|
| 898 |
+
},
|
| 899 |
+
{
|
| 900 |
+
"epoch": 2.87,
|
| 901 |
+
"learning_rate": 3.5673076923076926e-05,
|
| 902 |
+
"loss": 3.0173,
|
| 903 |
+
"step": 894
|
| 904 |
+
},
|
| 905 |
+
{
|
| 906 |
+
"epoch": 2.88,
|
| 907 |
+
"learning_rate": 3.557692307692308e-05,
|
| 908 |
+
"loss": 2.8656,
|
| 909 |
+
"step": 900
|
| 910 |
+
},
|
| 911 |
+
{
|
| 912 |
+
"epoch": 2.9,
|
| 913 |
+
"learning_rate": 3.548076923076924e-05,
|
| 914 |
+
"loss": 2.8834,
|
| 915 |
+
"step": 906
|
| 916 |
+
},
|
| 917 |
+
{
|
| 918 |
+
"epoch": 2.92,
|
| 919 |
+
"learning_rate": 3.538461538461539e-05,
|
| 920 |
+
"loss": 2.9829,
|
| 921 |
+
"step": 912
|
| 922 |
+
},
|
| 923 |
+
{
|
| 924 |
+
"epoch": 2.94,
|
| 925 |
+
"learning_rate": 3.528846153846154e-05,
|
| 926 |
+
"loss": 3.001,
|
| 927 |
+
"step": 918
|
| 928 |
+
},
|
| 929 |
+
{
|
| 930 |
+
"epoch": 2.96,
|
| 931 |
+
"learning_rate": 3.51923076923077e-05,
|
| 932 |
+
"loss": 2.9618,
|
| 933 |
+
"step": 924
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 2.98,
|
| 937 |
+
"learning_rate": 3.5096153846153845e-05,
|
| 938 |
+
"loss": 2.7964,
|
| 939 |
+
"step": 930
|
| 940 |
+
},
|
| 941 |
+
{
|
| 942 |
+
"epoch": 3.0,
|
| 943 |
+
"learning_rate": 3.5e-05,
|
| 944 |
+
"loss": 2.89,
|
| 945 |
+
"step": 936
|
| 946 |
+
},
|
| 947 |
+
{
|
| 948 |
+
"epoch": 3.02,
|
| 949 |
+
"learning_rate": 3.4903846153846156e-05,
|
| 950 |
+
"loss": 2.7191,
|
| 951 |
+
"step": 942
|
| 952 |
+
},
|
| 953 |
+
{
|
| 954 |
+
"epoch": 3.04,
|
| 955 |
+
"learning_rate": 3.480769230769231e-05,
|
| 956 |
+
"loss": 2.7875,
|
| 957 |
+
"step": 948
|
| 958 |
+
},
|
| 959 |
+
{
|
| 960 |
+
"epoch": 3.06,
|
| 961 |
+
"learning_rate": 3.471153846153846e-05,
|
| 962 |
+
"loss": 2.8498,
|
| 963 |
+
"step": 954
|
| 964 |
+
},
|
| 965 |
+
{
|
| 966 |
+
"epoch": 3.08,
|
| 967 |
+
"learning_rate": 3.461538461538462e-05,
|
| 968 |
+
"loss": 2.7858,
|
| 969 |
+
"step": 960
|
| 970 |
+
},
|
| 971 |
+
{
|
| 972 |
+
"epoch": 3.1,
|
| 973 |
+
"learning_rate": 3.451923076923077e-05,
|
| 974 |
+
"loss": 2.7435,
|
| 975 |
+
"step": 966
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 3.12,
|
| 979 |
+
"learning_rate": 3.442307692307692e-05,
|
| 980 |
+
"loss": 2.8449,
|
| 981 |
+
"step": 972
|
| 982 |
+
},
|
| 983 |
+
{
|
| 984 |
+
"epoch": 3.13,
|
| 985 |
+
"learning_rate": 3.432692307692308e-05,
|
| 986 |
+
"loss": 2.7817,
|
| 987 |
+
"step": 978
|
| 988 |
+
},
|
| 989 |
+
{
|
| 990 |
+
"epoch": 3.15,
|
| 991 |
+
"learning_rate": 3.4230769230769234e-05,
|
| 992 |
+
"loss": 2.8797,
|
| 993 |
+
"step": 984
|
| 994 |
+
},
|
| 995 |
+
{
|
| 996 |
+
"epoch": 3.17,
|
| 997 |
+
"learning_rate": 3.4134615384615386e-05,
|
| 998 |
+
"loss": 2.7719,
|
| 999 |
+
"step": 990
|
| 1000 |
+
},
|
| 1001 |
+
{
|
| 1002 |
+
"epoch": 3.19,
|
| 1003 |
+
"learning_rate": 3.4038461538461544e-05,
|
| 1004 |
+
"loss": 2.7919,
|
| 1005 |
+
"step": 996
|
| 1006 |
+
}
|
| 1007 |
+
],
|
| 1008 |
+
"logging_steps": 6,
|
| 1009 |
+
"max_steps": 3120,
|
| 1010 |
+
"num_train_epochs": 10,
|
| 1011 |
+
"save_steps": 500,
|
| 1012 |
+
"total_flos": 402675517440000.0,
|
| 1013 |
+
"trial_name": null,
|
| 1014 |
+
"trial_params": null
|
| 1015 |
+
}
|
checkpoint-1000/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c150481b2be1a59fc2f8a97e3f83518bf0fe89d39fbfc7d3d90a4bdd3f30ccaf
|
| 3 |
+
size 4600
|
checkpoint-1500/config.json
ADDED
|
@@ -0,0 +1,40 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "flax-community/gpt2-small-indonesian",
|
| 3 |
+
"activation_function": "gelu_new",
|
| 4 |
+
"architectures": [
|
| 5 |
+
"GPT2LMHeadModel"
|
| 6 |
+
],
|
| 7 |
+
"attn_pdrop": 0.0,
|
| 8 |
+
"bos_token_id": 50256,
|
| 9 |
+
"embd_pdrop": 0.0,
|
| 10 |
+
"eos_token_id": 50256,
|
| 11 |
+
"gradient_checkpointing": false,
|
| 12 |
+
"initializer_range": 0.02,
|
| 13 |
+
"layer_norm_epsilon": 1e-05,
|
| 14 |
+
"model_type": "gpt2",
|
| 15 |
+
"n_ctx": 1024,
|
| 16 |
+
"n_embd": 768,
|
| 17 |
+
"n_head": 12,
|
| 18 |
+
"n_inner": null,
|
| 19 |
+
"n_layer": 12,
|
| 20 |
+
"n_positions": 1024,
|
| 21 |
+
"reorder_and_upcast_attn": false,
|
| 22 |
+
"resid_pdrop": 0.0,
|
| 23 |
+
"scale_attn_by_inverse_layer_idx": false,
|
| 24 |
+
"scale_attn_weights": true,
|
| 25 |
+
"summary_activation": null,
|
| 26 |
+
"summary_first_dropout": 0.1,
|
| 27 |
+
"summary_proj_to_labels": true,
|
| 28 |
+
"summary_type": "cls_index",
|
| 29 |
+
"summary_use_proj": true,
|
| 30 |
+
"task_specific_params": {
|
| 31 |
+
"text-generation": {
|
| 32 |
+
"do_sample": true,
|
| 33 |
+
"max_length": 50
|
| 34 |
+
}
|
| 35 |
+
},
|
| 36 |
+
"torch_dtype": "float32",
|
| 37 |
+
"transformers_version": "4.35.2",
|
| 38 |
+
"use_cache": true,
|
| 39 |
+
"vocab_size": 50260
|
| 40 |
+
}
|
checkpoint-1500/generation_config.json
ADDED
|
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"bos_token_id": 50256,
|
| 4 |
+
"eos_token_id": 50256,
|
| 5 |
+
"transformers_version": "4.35.2"
|
| 6 |
+
}
|
checkpoint-1500/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c59a6761dce9256586a6cbedb58520090419bbfbe2647e55de236223d1d1090c
|
| 3 |
+
size 497783424
|
checkpoint-1500/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a30baf3cbdeb2ab5231a532a50e874ffcb0d6486af99361b33322dcf795a8584
|
| 3 |
+
size 995660730
|
checkpoint-1500/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d245e05e72192c132e0f2edb6fdcae0c578c890f0fe912f17ec7b0bba2d38cc3
|
| 3 |
+
size 14244
|
checkpoint-1500/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:590952593e733f3e830c7be42c6131abe9c2e79ec4f95c4484ad50107cfe5dcd
|
| 3 |
+
size 1064
|
checkpoint-1500/trainer_state.json
ADDED
|
@@ -0,0 +1,1519 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 4.8076923076923075,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 1500,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.02,
|
| 13 |
+
"learning_rate": 4.9903846153846154e-05,
|
| 14 |
+
"loss": 5.8218,
|
| 15 |
+
"step": 6
|
| 16 |
+
},
|
| 17 |
+
{
|
| 18 |
+
"epoch": 0.04,
|
| 19 |
+
"learning_rate": 4.980769230769231e-05,
|
| 20 |
+
"loss": 5.2499,
|
| 21 |
+
"step": 12
|
| 22 |
+
},
|
| 23 |
+
{
|
| 24 |
+
"epoch": 0.06,
|
| 25 |
+
"learning_rate": 4.9711538461538465e-05,
|
| 26 |
+
"loss": 5.1236,
|
| 27 |
+
"step": 18
|
| 28 |
+
},
|
| 29 |
+
{
|
| 30 |
+
"epoch": 0.08,
|
| 31 |
+
"learning_rate": 4.961538461538462e-05,
|
| 32 |
+
"loss": 4.8363,
|
| 33 |
+
"step": 24
|
| 34 |
+
},
|
| 35 |
+
{
|
| 36 |
+
"epoch": 0.1,
|
| 37 |
+
"learning_rate": 4.9519230769230776e-05,
|
| 38 |
+
"loss": 4.5805,
|
| 39 |
+
"step": 30
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.12,
|
| 43 |
+
"learning_rate": 4.942307692307693e-05,
|
| 44 |
+
"loss": 4.5276,
|
| 45 |
+
"step": 36
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.13,
|
| 49 |
+
"learning_rate": 4.932692307692308e-05,
|
| 50 |
+
"loss": 4.3871,
|
| 51 |
+
"step": 42
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.15,
|
| 55 |
+
"learning_rate": 4.923076923076924e-05,
|
| 56 |
+
"loss": 4.2706,
|
| 57 |
+
"step": 48
|
| 58 |
+
},
|
| 59 |
+
{
|
| 60 |
+
"epoch": 0.17,
|
| 61 |
+
"learning_rate": 4.9134615384615384e-05,
|
| 62 |
+
"loss": 4.0906,
|
| 63 |
+
"step": 54
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 0.19,
|
| 67 |
+
"learning_rate": 4.9038461538461536e-05,
|
| 68 |
+
"loss": 4.1704,
|
| 69 |
+
"step": 60
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 0.21,
|
| 73 |
+
"learning_rate": 4.8942307692307695e-05,
|
| 74 |
+
"loss": 3.9014,
|
| 75 |
+
"step": 66
|
| 76 |
+
},
|
| 77 |
+
{
|
| 78 |
+
"epoch": 0.23,
|
| 79 |
+
"learning_rate": 4.884615384615385e-05,
|
| 80 |
+
"loss": 4.0338,
|
| 81 |
+
"step": 72
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"epoch": 0.25,
|
| 85 |
+
"learning_rate": 4.875e-05,
|
| 86 |
+
"loss": 4.011,
|
| 87 |
+
"step": 78
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"epoch": 0.27,
|
| 91 |
+
"learning_rate": 4.865384615384616e-05,
|
| 92 |
+
"loss": 3.8017,
|
| 93 |
+
"step": 84
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.29,
|
| 97 |
+
"learning_rate": 4.855769230769231e-05,
|
| 98 |
+
"loss": 3.819,
|
| 99 |
+
"step": 90
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"epoch": 0.31,
|
| 103 |
+
"learning_rate": 4.846153846153846e-05,
|
| 104 |
+
"loss": 3.8157,
|
| 105 |
+
"step": 96
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 0.33,
|
| 109 |
+
"learning_rate": 4.836538461538462e-05,
|
| 110 |
+
"loss": 3.7675,
|
| 111 |
+
"step": 102
|
| 112 |
+
},
|
| 113 |
+
{
|
| 114 |
+
"epoch": 0.35,
|
| 115 |
+
"learning_rate": 4.826923076923077e-05,
|
| 116 |
+
"loss": 3.756,
|
| 117 |
+
"step": 108
|
| 118 |
+
},
|
| 119 |
+
{
|
| 120 |
+
"epoch": 0.37,
|
| 121 |
+
"learning_rate": 4.8173076923076925e-05,
|
| 122 |
+
"loss": 3.7439,
|
| 123 |
+
"step": 114
|
| 124 |
+
},
|
| 125 |
+
{
|
| 126 |
+
"epoch": 0.38,
|
| 127 |
+
"learning_rate": 4.8076923076923084e-05,
|
| 128 |
+
"loss": 3.761,
|
| 129 |
+
"step": 120
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.4,
|
| 133 |
+
"learning_rate": 4.7980769230769236e-05,
|
| 134 |
+
"loss": 3.7318,
|
| 135 |
+
"step": 126
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.42,
|
| 139 |
+
"learning_rate": 4.788461538461539e-05,
|
| 140 |
+
"loss": 3.745,
|
| 141 |
+
"step": 132
|
| 142 |
+
},
|
| 143 |
+
{
|
| 144 |
+
"epoch": 0.44,
|
| 145 |
+
"learning_rate": 4.778846153846154e-05,
|
| 146 |
+
"loss": 3.784,
|
| 147 |
+
"step": 138
|
| 148 |
+
},
|
| 149 |
+
{
|
| 150 |
+
"epoch": 0.46,
|
| 151 |
+
"learning_rate": 4.76923076923077e-05,
|
| 152 |
+
"loss": 3.6114,
|
| 153 |
+
"step": 144
|
| 154 |
+
},
|
| 155 |
+
{
|
| 156 |
+
"epoch": 0.48,
|
| 157 |
+
"learning_rate": 4.7596153846153844e-05,
|
| 158 |
+
"loss": 3.637,
|
| 159 |
+
"step": 150
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 0.5,
|
| 163 |
+
"learning_rate": 4.75e-05,
|
| 164 |
+
"loss": 3.5881,
|
| 165 |
+
"step": 156
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 0.52,
|
| 169 |
+
"learning_rate": 4.7403846153846155e-05,
|
| 170 |
+
"loss": 3.5336,
|
| 171 |
+
"step": 162
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 0.54,
|
| 175 |
+
"learning_rate": 4.730769230769231e-05,
|
| 176 |
+
"loss": 3.5313,
|
| 177 |
+
"step": 168
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.56,
|
| 181 |
+
"learning_rate": 4.7211538461538465e-05,
|
| 182 |
+
"loss": 3.6357,
|
| 183 |
+
"step": 174
|
| 184 |
+
},
|
| 185 |
+
{
|
| 186 |
+
"epoch": 0.58,
|
| 187 |
+
"learning_rate": 4.711538461538462e-05,
|
| 188 |
+
"loss": 3.5818,
|
| 189 |
+
"step": 180
|
| 190 |
+
},
|
| 191 |
+
{
|
| 192 |
+
"epoch": 0.6,
|
| 193 |
+
"learning_rate": 4.701923076923077e-05,
|
| 194 |
+
"loss": 3.5017,
|
| 195 |
+
"step": 186
|
| 196 |
+
},
|
| 197 |
+
{
|
| 198 |
+
"epoch": 0.62,
|
| 199 |
+
"learning_rate": 4.692307692307693e-05,
|
| 200 |
+
"loss": 3.5815,
|
| 201 |
+
"step": 192
|
| 202 |
+
},
|
| 203 |
+
{
|
| 204 |
+
"epoch": 0.63,
|
| 205 |
+
"learning_rate": 4.682692307692308e-05,
|
| 206 |
+
"loss": 3.611,
|
| 207 |
+
"step": 198
|
| 208 |
+
},
|
| 209 |
+
{
|
| 210 |
+
"epoch": 0.65,
|
| 211 |
+
"learning_rate": 4.673076923076923e-05,
|
| 212 |
+
"loss": 3.6096,
|
| 213 |
+
"step": 204
|
| 214 |
+
},
|
| 215 |
+
{
|
| 216 |
+
"epoch": 0.67,
|
| 217 |
+
"learning_rate": 4.6634615384615384e-05,
|
| 218 |
+
"loss": 3.5436,
|
| 219 |
+
"step": 210
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.69,
|
| 223 |
+
"learning_rate": 4.653846153846154e-05,
|
| 224 |
+
"loss": 3.5258,
|
| 225 |
+
"step": 216
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 0.71,
|
| 229 |
+
"learning_rate": 4.6442307692307695e-05,
|
| 230 |
+
"loss": 3.5277,
|
| 231 |
+
"step": 222
|
| 232 |
+
},
|
| 233 |
+
{
|
| 234 |
+
"epoch": 0.73,
|
| 235 |
+
"learning_rate": 4.634615384615385e-05,
|
| 236 |
+
"loss": 3.505,
|
| 237 |
+
"step": 228
|
| 238 |
+
},
|
| 239 |
+
{
|
| 240 |
+
"epoch": 0.75,
|
| 241 |
+
"learning_rate": 4.6250000000000006e-05,
|
| 242 |
+
"loss": 3.4665,
|
| 243 |
+
"step": 234
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"epoch": 0.77,
|
| 247 |
+
"learning_rate": 4.615384615384616e-05,
|
| 248 |
+
"loss": 3.4245,
|
| 249 |
+
"step": 240
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 0.79,
|
| 253 |
+
"learning_rate": 4.605769230769231e-05,
|
| 254 |
+
"loss": 3.4547,
|
| 255 |
+
"step": 246
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 0.81,
|
| 259 |
+
"learning_rate": 4.596153846153846e-05,
|
| 260 |
+
"loss": 3.4382,
|
| 261 |
+
"step": 252
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.83,
|
| 265 |
+
"learning_rate": 4.5865384615384614e-05,
|
| 266 |
+
"loss": 3.482,
|
| 267 |
+
"step": 258
|
| 268 |
+
},
|
| 269 |
+
{
|
| 270 |
+
"epoch": 0.85,
|
| 271 |
+
"learning_rate": 4.576923076923077e-05,
|
| 272 |
+
"loss": 3.4795,
|
| 273 |
+
"step": 264
|
| 274 |
+
},
|
| 275 |
+
{
|
| 276 |
+
"epoch": 0.87,
|
| 277 |
+
"learning_rate": 4.5673076923076925e-05,
|
| 278 |
+
"loss": 3.5018,
|
| 279 |
+
"step": 270
|
| 280 |
+
},
|
| 281 |
+
{
|
| 282 |
+
"epoch": 0.88,
|
| 283 |
+
"learning_rate": 4.557692307692308e-05,
|
| 284 |
+
"loss": 3.448,
|
| 285 |
+
"step": 276
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 0.9,
|
| 289 |
+
"learning_rate": 4.548076923076923e-05,
|
| 290 |
+
"loss": 3.4206,
|
| 291 |
+
"step": 282
|
| 292 |
+
},
|
| 293 |
+
{
|
| 294 |
+
"epoch": 0.92,
|
| 295 |
+
"learning_rate": 4.538461538461539e-05,
|
| 296 |
+
"loss": 3.528,
|
| 297 |
+
"step": 288
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 0.94,
|
| 301 |
+
"learning_rate": 4.528846153846154e-05,
|
| 302 |
+
"loss": 3.475,
|
| 303 |
+
"step": 294
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.96,
|
| 307 |
+
"learning_rate": 4.519230769230769e-05,
|
| 308 |
+
"loss": 3.4142,
|
| 309 |
+
"step": 300
|
| 310 |
+
},
|
| 311 |
+
{
|
| 312 |
+
"epoch": 0.98,
|
| 313 |
+
"learning_rate": 4.509615384615385e-05,
|
| 314 |
+
"loss": 3.4156,
|
| 315 |
+
"step": 306
|
| 316 |
+
},
|
| 317 |
+
{
|
| 318 |
+
"epoch": 1.0,
|
| 319 |
+
"learning_rate": 4.5e-05,
|
| 320 |
+
"loss": 3.3274,
|
| 321 |
+
"step": 312
|
| 322 |
+
},
|
| 323 |
+
{
|
| 324 |
+
"epoch": 1.02,
|
| 325 |
+
"learning_rate": 4.4903846153846155e-05,
|
| 326 |
+
"loss": 3.3594,
|
| 327 |
+
"step": 318
|
| 328 |
+
},
|
| 329 |
+
{
|
| 330 |
+
"epoch": 1.04,
|
| 331 |
+
"learning_rate": 4.4807692307692314e-05,
|
| 332 |
+
"loss": 3.2517,
|
| 333 |
+
"step": 324
|
| 334 |
+
},
|
| 335 |
+
{
|
| 336 |
+
"epoch": 1.06,
|
| 337 |
+
"learning_rate": 4.4711538461538466e-05,
|
| 338 |
+
"loss": 3.2934,
|
| 339 |
+
"step": 330
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 1.08,
|
| 343 |
+
"learning_rate": 4.461538461538462e-05,
|
| 344 |
+
"loss": 3.2986,
|
| 345 |
+
"step": 336
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 1.1,
|
| 349 |
+
"learning_rate": 4.451923076923077e-05,
|
| 350 |
+
"loss": 3.2655,
|
| 351 |
+
"step": 342
|
| 352 |
+
},
|
| 353 |
+
{
|
| 354 |
+
"epoch": 1.12,
|
| 355 |
+
"learning_rate": 4.442307692307692e-05,
|
| 356 |
+
"loss": 3.4058,
|
| 357 |
+
"step": 348
|
| 358 |
+
},
|
| 359 |
+
{
|
| 360 |
+
"epoch": 1.13,
|
| 361 |
+
"learning_rate": 4.4326923076923074e-05,
|
| 362 |
+
"loss": 3.265,
|
| 363 |
+
"step": 354
|
| 364 |
+
},
|
| 365 |
+
{
|
| 366 |
+
"epoch": 1.15,
|
| 367 |
+
"learning_rate": 4.423076923076923e-05,
|
| 368 |
+
"loss": 3.3208,
|
| 369 |
+
"step": 360
|
| 370 |
+
},
|
| 371 |
+
{
|
| 372 |
+
"epoch": 1.17,
|
| 373 |
+
"learning_rate": 4.4134615384615385e-05,
|
| 374 |
+
"loss": 3.238,
|
| 375 |
+
"step": 366
|
| 376 |
+
},
|
| 377 |
+
{
|
| 378 |
+
"epoch": 1.19,
|
| 379 |
+
"learning_rate": 4.403846153846154e-05,
|
| 380 |
+
"loss": 3.2153,
|
| 381 |
+
"step": 372
|
| 382 |
+
},
|
| 383 |
+
{
|
| 384 |
+
"epoch": 1.21,
|
| 385 |
+
"learning_rate": 4.3942307692307695e-05,
|
| 386 |
+
"loss": 3.264,
|
| 387 |
+
"step": 378
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 1.23,
|
| 391 |
+
"learning_rate": 4.384615384615385e-05,
|
| 392 |
+
"loss": 3.2911,
|
| 393 |
+
"step": 384
|
| 394 |
+
},
|
| 395 |
+
{
|
| 396 |
+
"epoch": 1.25,
|
| 397 |
+
"learning_rate": 4.375e-05,
|
| 398 |
+
"loss": 3.3027,
|
| 399 |
+
"step": 390
|
| 400 |
+
},
|
| 401 |
+
{
|
| 402 |
+
"epoch": 1.27,
|
| 403 |
+
"learning_rate": 4.365384615384616e-05,
|
| 404 |
+
"loss": 3.2589,
|
| 405 |
+
"step": 396
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 1.29,
|
| 409 |
+
"learning_rate": 4.355769230769231e-05,
|
| 410 |
+
"loss": 3.3683,
|
| 411 |
+
"step": 402
|
| 412 |
+
},
|
| 413 |
+
{
|
| 414 |
+
"epoch": 1.31,
|
| 415 |
+
"learning_rate": 4.346153846153846e-05,
|
| 416 |
+
"loss": 3.2849,
|
| 417 |
+
"step": 408
|
| 418 |
+
},
|
| 419 |
+
{
|
| 420 |
+
"epoch": 1.33,
|
| 421 |
+
"learning_rate": 4.336538461538462e-05,
|
| 422 |
+
"loss": 3.2397,
|
| 423 |
+
"step": 414
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 1.35,
|
| 427 |
+
"learning_rate": 4.326923076923077e-05,
|
| 428 |
+
"loss": 3.2128,
|
| 429 |
+
"step": 420
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 1.37,
|
| 433 |
+
"learning_rate": 4.3173076923076925e-05,
|
| 434 |
+
"loss": 3.1944,
|
| 435 |
+
"step": 426
|
| 436 |
+
},
|
| 437 |
+
{
|
| 438 |
+
"epoch": 1.38,
|
| 439 |
+
"learning_rate": 4.3076923076923084e-05,
|
| 440 |
+
"loss": 3.1837,
|
| 441 |
+
"step": 432
|
| 442 |
+
},
|
| 443 |
+
{
|
| 444 |
+
"epoch": 1.4,
|
| 445 |
+
"learning_rate": 4.2980769230769236e-05,
|
| 446 |
+
"loss": 3.1793,
|
| 447 |
+
"step": 438
|
| 448 |
+
},
|
| 449 |
+
{
|
| 450 |
+
"epoch": 1.42,
|
| 451 |
+
"learning_rate": 4.288461538461538e-05,
|
| 452 |
+
"loss": 3.1447,
|
| 453 |
+
"step": 444
|
| 454 |
+
},
|
| 455 |
+
{
|
| 456 |
+
"epoch": 1.44,
|
| 457 |
+
"learning_rate": 4.278846153846154e-05,
|
| 458 |
+
"loss": 3.1028,
|
| 459 |
+
"step": 450
|
| 460 |
+
},
|
| 461 |
+
{
|
| 462 |
+
"epoch": 1.46,
|
| 463 |
+
"learning_rate": 4.269230769230769e-05,
|
| 464 |
+
"loss": 3.2471,
|
| 465 |
+
"step": 456
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 1.48,
|
| 469 |
+
"learning_rate": 4.2596153846153844e-05,
|
| 470 |
+
"loss": 3.1855,
|
| 471 |
+
"step": 462
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 1.5,
|
| 475 |
+
"learning_rate": 4.25e-05,
|
| 476 |
+
"loss": 3.1817,
|
| 477 |
+
"step": 468
|
| 478 |
+
},
|
| 479 |
+
{
|
| 480 |
+
"epoch": 1.52,
|
| 481 |
+
"learning_rate": 4.2403846153846155e-05,
|
| 482 |
+
"loss": 3.214,
|
| 483 |
+
"step": 474
|
| 484 |
+
},
|
| 485 |
+
{
|
| 486 |
+
"epoch": 1.54,
|
| 487 |
+
"learning_rate": 4.230769230769231e-05,
|
| 488 |
+
"loss": 2.997,
|
| 489 |
+
"step": 480
|
| 490 |
+
},
|
| 491 |
+
{
|
| 492 |
+
"epoch": 1.56,
|
| 493 |
+
"learning_rate": 4.2211538461538466e-05,
|
| 494 |
+
"loss": 3.2059,
|
| 495 |
+
"step": 486
|
| 496 |
+
},
|
| 497 |
+
{
|
| 498 |
+
"epoch": 1.58,
|
| 499 |
+
"learning_rate": 4.211538461538462e-05,
|
| 500 |
+
"loss": 3.1517,
|
| 501 |
+
"step": 492
|
| 502 |
+
},
|
| 503 |
+
{
|
| 504 |
+
"epoch": 1.6,
|
| 505 |
+
"learning_rate": 4.201923076923077e-05,
|
| 506 |
+
"loss": 3.2236,
|
| 507 |
+
"step": 498
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 1.62,
|
| 511 |
+
"learning_rate": 4.192307692307693e-05,
|
| 512 |
+
"loss": 3.1939,
|
| 513 |
+
"step": 504
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 1.63,
|
| 517 |
+
"learning_rate": 4.182692307692308e-05,
|
| 518 |
+
"loss": 3.1143,
|
| 519 |
+
"step": 510
|
| 520 |
+
},
|
| 521 |
+
{
|
| 522 |
+
"epoch": 1.65,
|
| 523 |
+
"learning_rate": 4.173076923076923e-05,
|
| 524 |
+
"loss": 3.2261,
|
| 525 |
+
"step": 516
|
| 526 |
+
},
|
| 527 |
+
{
|
| 528 |
+
"epoch": 1.67,
|
| 529 |
+
"learning_rate": 4.163461538461539e-05,
|
| 530 |
+
"loss": 3.2095,
|
| 531 |
+
"step": 522
|
| 532 |
+
},
|
| 533 |
+
{
|
| 534 |
+
"epoch": 1.69,
|
| 535 |
+
"learning_rate": 4.1538461538461544e-05,
|
| 536 |
+
"loss": 3.2498,
|
| 537 |
+
"step": 528
|
| 538 |
+
},
|
| 539 |
+
{
|
| 540 |
+
"epoch": 1.71,
|
| 541 |
+
"learning_rate": 4.1442307692307696e-05,
|
| 542 |
+
"loss": 3.1493,
|
| 543 |
+
"step": 534
|
| 544 |
+
},
|
| 545 |
+
{
|
| 546 |
+
"epoch": 1.73,
|
| 547 |
+
"learning_rate": 4.134615384615385e-05,
|
| 548 |
+
"loss": 3.0919,
|
| 549 |
+
"step": 540
|
| 550 |
+
},
|
| 551 |
+
{
|
| 552 |
+
"epoch": 1.75,
|
| 553 |
+
"learning_rate": 4.125e-05,
|
| 554 |
+
"loss": 3.2617,
|
| 555 |
+
"step": 546
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 1.77,
|
| 559 |
+
"learning_rate": 4.115384615384615e-05,
|
| 560 |
+
"loss": 3.1447,
|
| 561 |
+
"step": 552
|
| 562 |
+
},
|
| 563 |
+
{
|
| 564 |
+
"epoch": 1.79,
|
| 565 |
+
"learning_rate": 4.105769230769231e-05,
|
| 566 |
+
"loss": 3.2025,
|
| 567 |
+
"step": 558
|
| 568 |
+
},
|
| 569 |
+
{
|
| 570 |
+
"epoch": 1.81,
|
| 571 |
+
"learning_rate": 4.096153846153846e-05,
|
| 572 |
+
"loss": 3.1829,
|
| 573 |
+
"step": 564
|
| 574 |
+
},
|
| 575 |
+
{
|
| 576 |
+
"epoch": 1.83,
|
| 577 |
+
"learning_rate": 4.0865384615384615e-05,
|
| 578 |
+
"loss": 3.2216,
|
| 579 |
+
"step": 570
|
| 580 |
+
},
|
| 581 |
+
{
|
| 582 |
+
"epoch": 1.85,
|
| 583 |
+
"learning_rate": 4.0769230769230773e-05,
|
| 584 |
+
"loss": 3.1391,
|
| 585 |
+
"step": 576
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 1.87,
|
| 589 |
+
"learning_rate": 4.0673076923076926e-05,
|
| 590 |
+
"loss": 3.1344,
|
| 591 |
+
"step": 582
|
| 592 |
+
},
|
| 593 |
+
{
|
| 594 |
+
"epoch": 1.88,
|
| 595 |
+
"learning_rate": 4.057692307692308e-05,
|
| 596 |
+
"loss": 3.1699,
|
| 597 |
+
"step": 588
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 1.9,
|
| 601 |
+
"learning_rate": 4.0480769230769236e-05,
|
| 602 |
+
"loss": 3.0138,
|
| 603 |
+
"step": 594
|
| 604 |
+
},
|
| 605 |
+
{
|
| 606 |
+
"epoch": 1.92,
|
| 607 |
+
"learning_rate": 4.038461538461539e-05,
|
| 608 |
+
"loss": 3.1157,
|
| 609 |
+
"step": 600
|
| 610 |
+
},
|
| 611 |
+
{
|
| 612 |
+
"epoch": 1.94,
|
| 613 |
+
"learning_rate": 4.028846153846154e-05,
|
| 614 |
+
"loss": 3.1173,
|
| 615 |
+
"step": 606
|
| 616 |
+
},
|
| 617 |
+
{
|
| 618 |
+
"epoch": 1.96,
|
| 619 |
+
"learning_rate": 4.019230769230769e-05,
|
| 620 |
+
"loss": 3.2019,
|
| 621 |
+
"step": 612
|
| 622 |
+
},
|
| 623 |
+
{
|
| 624 |
+
"epoch": 1.98,
|
| 625 |
+
"learning_rate": 4.009615384615385e-05,
|
| 626 |
+
"loss": 3.0801,
|
| 627 |
+
"step": 618
|
| 628 |
+
},
|
| 629 |
+
{
|
| 630 |
+
"epoch": 2.0,
|
| 631 |
+
"learning_rate": 4e-05,
|
| 632 |
+
"loss": 3.0888,
|
| 633 |
+
"step": 624
|
| 634 |
+
},
|
| 635 |
+
{
|
| 636 |
+
"epoch": 2.02,
|
| 637 |
+
"learning_rate": 3.9903846153846155e-05,
|
| 638 |
+
"loss": 3.0332,
|
| 639 |
+
"step": 630
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 2.04,
|
| 643 |
+
"learning_rate": 3.980769230769231e-05,
|
| 644 |
+
"loss": 2.9418,
|
| 645 |
+
"step": 636
|
| 646 |
+
},
|
| 647 |
+
{
|
| 648 |
+
"epoch": 2.06,
|
| 649 |
+
"learning_rate": 3.971153846153846e-05,
|
| 650 |
+
"loss": 2.933,
|
| 651 |
+
"step": 642
|
| 652 |
+
},
|
| 653 |
+
{
|
| 654 |
+
"epoch": 2.08,
|
| 655 |
+
"learning_rate": 3.961538461538462e-05,
|
| 656 |
+
"loss": 3.0467,
|
| 657 |
+
"step": 648
|
| 658 |
+
},
|
| 659 |
+
{
|
| 660 |
+
"epoch": 2.1,
|
| 661 |
+
"learning_rate": 3.951923076923077e-05,
|
| 662 |
+
"loss": 3.0991,
|
| 663 |
+
"step": 654
|
| 664 |
+
},
|
| 665 |
+
{
|
| 666 |
+
"epoch": 2.12,
|
| 667 |
+
"learning_rate": 3.942307692307692e-05,
|
| 668 |
+
"loss": 2.913,
|
| 669 |
+
"step": 660
|
| 670 |
+
},
|
| 671 |
+
{
|
| 672 |
+
"epoch": 2.13,
|
| 673 |
+
"learning_rate": 3.932692307692308e-05,
|
| 674 |
+
"loss": 3.0531,
|
| 675 |
+
"step": 666
|
| 676 |
+
},
|
| 677 |
+
{
|
| 678 |
+
"epoch": 2.15,
|
| 679 |
+
"learning_rate": 3.923076923076923e-05,
|
| 680 |
+
"loss": 2.9838,
|
| 681 |
+
"step": 672
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 2.17,
|
| 685 |
+
"learning_rate": 3.9134615384615385e-05,
|
| 686 |
+
"loss": 2.9406,
|
| 687 |
+
"step": 678
|
| 688 |
+
},
|
| 689 |
+
{
|
| 690 |
+
"epoch": 2.19,
|
| 691 |
+
"learning_rate": 3.903846153846154e-05,
|
| 692 |
+
"loss": 3.046,
|
| 693 |
+
"step": 684
|
| 694 |
+
},
|
| 695 |
+
{
|
| 696 |
+
"epoch": 2.21,
|
| 697 |
+
"learning_rate": 3.8942307692307696e-05,
|
| 698 |
+
"loss": 3.006,
|
| 699 |
+
"step": 690
|
| 700 |
+
},
|
| 701 |
+
{
|
| 702 |
+
"epoch": 2.23,
|
| 703 |
+
"learning_rate": 3.884615384615385e-05,
|
| 704 |
+
"loss": 2.9774,
|
| 705 |
+
"step": 696
|
| 706 |
+
},
|
| 707 |
+
{
|
| 708 |
+
"epoch": 2.25,
|
| 709 |
+
"learning_rate": 3.875e-05,
|
| 710 |
+
"loss": 2.9937,
|
| 711 |
+
"step": 702
|
| 712 |
+
},
|
| 713 |
+
{
|
| 714 |
+
"epoch": 2.27,
|
| 715 |
+
"learning_rate": 3.865384615384616e-05,
|
| 716 |
+
"loss": 2.9737,
|
| 717 |
+
"step": 708
|
| 718 |
+
},
|
| 719 |
+
{
|
| 720 |
+
"epoch": 2.29,
|
| 721 |
+
"learning_rate": 3.855769230769231e-05,
|
| 722 |
+
"loss": 2.9772,
|
| 723 |
+
"step": 714
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 2.31,
|
| 727 |
+
"learning_rate": 3.846153846153846e-05,
|
| 728 |
+
"loss": 2.8405,
|
| 729 |
+
"step": 720
|
| 730 |
+
},
|
| 731 |
+
{
|
| 732 |
+
"epoch": 2.33,
|
| 733 |
+
"learning_rate": 3.836538461538462e-05,
|
| 734 |
+
"loss": 3.0856,
|
| 735 |
+
"step": 726
|
| 736 |
+
},
|
| 737 |
+
{
|
| 738 |
+
"epoch": 2.35,
|
| 739 |
+
"learning_rate": 3.826923076923077e-05,
|
| 740 |
+
"loss": 2.9947,
|
| 741 |
+
"step": 732
|
| 742 |
+
},
|
| 743 |
+
{
|
| 744 |
+
"epoch": 2.37,
|
| 745 |
+
"learning_rate": 3.8173076923076926e-05,
|
| 746 |
+
"loss": 2.9251,
|
| 747 |
+
"step": 738
|
| 748 |
+
},
|
| 749 |
+
{
|
| 750 |
+
"epoch": 2.38,
|
| 751 |
+
"learning_rate": 3.807692307692308e-05,
|
| 752 |
+
"loss": 2.9613,
|
| 753 |
+
"step": 744
|
| 754 |
+
},
|
| 755 |
+
{
|
| 756 |
+
"epoch": 2.4,
|
| 757 |
+
"learning_rate": 3.798076923076923e-05,
|
| 758 |
+
"loss": 2.9402,
|
| 759 |
+
"step": 750
|
| 760 |
+
},
|
| 761 |
+
{
|
| 762 |
+
"epoch": 2.42,
|
| 763 |
+
"learning_rate": 3.788461538461538e-05,
|
| 764 |
+
"loss": 2.8823,
|
| 765 |
+
"step": 756
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 2.44,
|
| 769 |
+
"learning_rate": 3.778846153846154e-05,
|
| 770 |
+
"loss": 2.8906,
|
| 771 |
+
"step": 762
|
| 772 |
+
},
|
| 773 |
+
{
|
| 774 |
+
"epoch": 2.46,
|
| 775 |
+
"learning_rate": 3.769230769230769e-05,
|
| 776 |
+
"loss": 2.9475,
|
| 777 |
+
"step": 768
|
| 778 |
+
},
|
| 779 |
+
{
|
| 780 |
+
"epoch": 2.48,
|
| 781 |
+
"learning_rate": 3.7596153846153845e-05,
|
| 782 |
+
"loss": 2.9999,
|
| 783 |
+
"step": 774
|
| 784 |
+
},
|
| 785 |
+
{
|
| 786 |
+
"epoch": 2.5,
|
| 787 |
+
"learning_rate": 3.7500000000000003e-05,
|
| 788 |
+
"loss": 2.9503,
|
| 789 |
+
"step": 780
|
| 790 |
+
},
|
| 791 |
+
{
|
| 792 |
+
"epoch": 2.52,
|
| 793 |
+
"learning_rate": 3.7403846153846156e-05,
|
| 794 |
+
"loss": 3.0112,
|
| 795 |
+
"step": 786
|
| 796 |
+
},
|
| 797 |
+
{
|
| 798 |
+
"epoch": 2.54,
|
| 799 |
+
"learning_rate": 3.730769230769231e-05,
|
| 800 |
+
"loss": 3.0096,
|
| 801 |
+
"step": 792
|
| 802 |
+
},
|
| 803 |
+
{
|
| 804 |
+
"epoch": 2.56,
|
| 805 |
+
"learning_rate": 3.7211538461538466e-05,
|
| 806 |
+
"loss": 2.9804,
|
| 807 |
+
"step": 798
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 2.58,
|
| 811 |
+
"learning_rate": 3.711538461538462e-05,
|
| 812 |
+
"loss": 2.9203,
|
| 813 |
+
"step": 804
|
| 814 |
+
},
|
| 815 |
+
{
|
| 816 |
+
"epoch": 2.6,
|
| 817 |
+
"learning_rate": 3.701923076923077e-05,
|
| 818 |
+
"loss": 2.939,
|
| 819 |
+
"step": 810
|
| 820 |
+
},
|
| 821 |
+
{
|
| 822 |
+
"epoch": 2.62,
|
| 823 |
+
"learning_rate": 3.692307692307693e-05,
|
| 824 |
+
"loss": 2.8455,
|
| 825 |
+
"step": 816
|
| 826 |
+
},
|
| 827 |
+
{
|
| 828 |
+
"epoch": 2.63,
|
| 829 |
+
"learning_rate": 3.682692307692308e-05,
|
| 830 |
+
"loss": 2.9651,
|
| 831 |
+
"step": 822
|
| 832 |
+
},
|
| 833 |
+
{
|
| 834 |
+
"epoch": 2.65,
|
| 835 |
+
"learning_rate": 3.673076923076923e-05,
|
| 836 |
+
"loss": 2.9528,
|
| 837 |
+
"step": 828
|
| 838 |
+
},
|
| 839 |
+
{
|
| 840 |
+
"epoch": 2.67,
|
| 841 |
+
"learning_rate": 3.6634615384615385e-05,
|
| 842 |
+
"loss": 2.8042,
|
| 843 |
+
"step": 834
|
| 844 |
+
},
|
| 845 |
+
{
|
| 846 |
+
"epoch": 2.69,
|
| 847 |
+
"learning_rate": 3.653846153846154e-05,
|
| 848 |
+
"loss": 2.8311,
|
| 849 |
+
"step": 840
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 2.71,
|
| 853 |
+
"learning_rate": 3.644230769230769e-05,
|
| 854 |
+
"loss": 2.8888,
|
| 855 |
+
"step": 846
|
| 856 |
+
},
|
| 857 |
+
{
|
| 858 |
+
"epoch": 2.73,
|
| 859 |
+
"learning_rate": 3.634615384615385e-05,
|
| 860 |
+
"loss": 2.9151,
|
| 861 |
+
"step": 852
|
| 862 |
+
},
|
| 863 |
+
{
|
| 864 |
+
"epoch": 2.75,
|
| 865 |
+
"learning_rate": 3.625e-05,
|
| 866 |
+
"loss": 2.9463,
|
| 867 |
+
"step": 858
|
| 868 |
+
},
|
| 869 |
+
{
|
| 870 |
+
"epoch": 2.77,
|
| 871 |
+
"learning_rate": 3.615384615384615e-05,
|
| 872 |
+
"loss": 2.957,
|
| 873 |
+
"step": 864
|
| 874 |
+
},
|
| 875 |
+
{
|
| 876 |
+
"epoch": 2.79,
|
| 877 |
+
"learning_rate": 3.605769230769231e-05,
|
| 878 |
+
"loss": 2.9473,
|
| 879 |
+
"step": 870
|
| 880 |
+
},
|
| 881 |
+
{
|
| 882 |
+
"epoch": 2.81,
|
| 883 |
+
"learning_rate": 3.596153846153846e-05,
|
| 884 |
+
"loss": 2.9994,
|
| 885 |
+
"step": 876
|
| 886 |
+
},
|
| 887 |
+
{
|
| 888 |
+
"epoch": 2.83,
|
| 889 |
+
"learning_rate": 3.5865384615384615e-05,
|
| 890 |
+
"loss": 3.0486,
|
| 891 |
+
"step": 882
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 2.85,
|
| 895 |
+
"learning_rate": 3.5769230769230774e-05,
|
| 896 |
+
"loss": 2.9487,
|
| 897 |
+
"step": 888
|
| 898 |
+
},
|
| 899 |
+
{
|
| 900 |
+
"epoch": 2.87,
|
| 901 |
+
"learning_rate": 3.5673076923076926e-05,
|
| 902 |
+
"loss": 3.0173,
|
| 903 |
+
"step": 894
|
| 904 |
+
},
|
| 905 |
+
{
|
| 906 |
+
"epoch": 2.88,
|
| 907 |
+
"learning_rate": 3.557692307692308e-05,
|
| 908 |
+
"loss": 2.8656,
|
| 909 |
+
"step": 900
|
| 910 |
+
},
|
| 911 |
+
{
|
| 912 |
+
"epoch": 2.9,
|
| 913 |
+
"learning_rate": 3.548076923076924e-05,
|
| 914 |
+
"loss": 2.8834,
|
| 915 |
+
"step": 906
|
| 916 |
+
},
|
| 917 |
+
{
|
| 918 |
+
"epoch": 2.92,
|
| 919 |
+
"learning_rate": 3.538461538461539e-05,
|
| 920 |
+
"loss": 2.9829,
|
| 921 |
+
"step": 912
|
| 922 |
+
},
|
| 923 |
+
{
|
| 924 |
+
"epoch": 2.94,
|
| 925 |
+
"learning_rate": 3.528846153846154e-05,
|
| 926 |
+
"loss": 3.001,
|
| 927 |
+
"step": 918
|
| 928 |
+
},
|
| 929 |
+
{
|
| 930 |
+
"epoch": 2.96,
|
| 931 |
+
"learning_rate": 3.51923076923077e-05,
|
| 932 |
+
"loss": 2.9618,
|
| 933 |
+
"step": 924
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 2.98,
|
| 937 |
+
"learning_rate": 3.5096153846153845e-05,
|
| 938 |
+
"loss": 2.7964,
|
| 939 |
+
"step": 930
|
| 940 |
+
},
|
| 941 |
+
{
|
| 942 |
+
"epoch": 3.0,
|
| 943 |
+
"learning_rate": 3.5e-05,
|
| 944 |
+
"loss": 2.89,
|
| 945 |
+
"step": 936
|
| 946 |
+
},
|
| 947 |
+
{
|
| 948 |
+
"epoch": 3.02,
|
| 949 |
+
"learning_rate": 3.4903846153846156e-05,
|
| 950 |
+
"loss": 2.7191,
|
| 951 |
+
"step": 942
|
| 952 |
+
},
|
| 953 |
+
{
|
| 954 |
+
"epoch": 3.04,
|
| 955 |
+
"learning_rate": 3.480769230769231e-05,
|
| 956 |
+
"loss": 2.7875,
|
| 957 |
+
"step": 948
|
| 958 |
+
},
|
| 959 |
+
{
|
| 960 |
+
"epoch": 3.06,
|
| 961 |
+
"learning_rate": 3.471153846153846e-05,
|
| 962 |
+
"loss": 2.8498,
|
| 963 |
+
"step": 954
|
| 964 |
+
},
|
| 965 |
+
{
|
| 966 |
+
"epoch": 3.08,
|
| 967 |
+
"learning_rate": 3.461538461538462e-05,
|
| 968 |
+
"loss": 2.7858,
|
| 969 |
+
"step": 960
|
| 970 |
+
},
|
| 971 |
+
{
|
| 972 |
+
"epoch": 3.1,
|
| 973 |
+
"learning_rate": 3.451923076923077e-05,
|
| 974 |
+
"loss": 2.7435,
|
| 975 |
+
"step": 966
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 3.12,
|
| 979 |
+
"learning_rate": 3.442307692307692e-05,
|
| 980 |
+
"loss": 2.8449,
|
| 981 |
+
"step": 972
|
| 982 |
+
},
|
| 983 |
+
{
|
| 984 |
+
"epoch": 3.13,
|
| 985 |
+
"learning_rate": 3.432692307692308e-05,
|
| 986 |
+
"loss": 2.7817,
|
| 987 |
+
"step": 978
|
| 988 |
+
},
|
| 989 |
+
{
|
| 990 |
+
"epoch": 3.15,
|
| 991 |
+
"learning_rate": 3.4230769230769234e-05,
|
| 992 |
+
"loss": 2.8797,
|
| 993 |
+
"step": 984
|
| 994 |
+
},
|
| 995 |
+
{
|
| 996 |
+
"epoch": 3.17,
|
| 997 |
+
"learning_rate": 3.4134615384615386e-05,
|
| 998 |
+
"loss": 2.7719,
|
| 999 |
+
"step": 990
|
| 1000 |
+
},
|
| 1001 |
+
{
|
| 1002 |
+
"epoch": 3.19,
|
| 1003 |
+
"learning_rate": 3.4038461538461544e-05,
|
| 1004 |
+
"loss": 2.7919,
|
| 1005 |
+
"step": 996
|
| 1006 |
+
},
|
| 1007 |
+
{
|
| 1008 |
+
"epoch": 3.21,
|
| 1009 |
+
"learning_rate": 3.3942307692307696e-05,
|
| 1010 |
+
"loss": 2.8072,
|
| 1011 |
+
"step": 1002
|
| 1012 |
+
},
|
| 1013 |
+
{
|
| 1014 |
+
"epoch": 3.23,
|
| 1015 |
+
"learning_rate": 3.384615384615385e-05,
|
| 1016 |
+
"loss": 2.7281,
|
| 1017 |
+
"step": 1008
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 3.25,
|
| 1021 |
+
"learning_rate": 3.375000000000001e-05,
|
| 1022 |
+
"loss": 2.8431,
|
| 1023 |
+
"step": 1014
|
| 1024 |
+
},
|
| 1025 |
+
{
|
| 1026 |
+
"epoch": 3.27,
|
| 1027 |
+
"learning_rate": 3.365384615384616e-05,
|
| 1028 |
+
"loss": 2.7154,
|
| 1029 |
+
"step": 1020
|
| 1030 |
+
},
|
| 1031 |
+
{
|
| 1032 |
+
"epoch": 3.29,
|
| 1033 |
+
"learning_rate": 3.3557692307692304e-05,
|
| 1034 |
+
"loss": 2.83,
|
| 1035 |
+
"step": 1026
|
| 1036 |
+
},
|
| 1037 |
+
{
|
| 1038 |
+
"epoch": 3.31,
|
| 1039 |
+
"learning_rate": 3.346153846153846e-05,
|
| 1040 |
+
"loss": 2.8434,
|
| 1041 |
+
"step": 1032
|
| 1042 |
+
},
|
| 1043 |
+
{
|
| 1044 |
+
"epoch": 3.33,
|
| 1045 |
+
"learning_rate": 3.3365384615384615e-05,
|
| 1046 |
+
"loss": 2.8399,
|
| 1047 |
+
"step": 1038
|
| 1048 |
+
},
|
| 1049 |
+
{
|
| 1050 |
+
"epoch": 3.35,
|
| 1051 |
+
"learning_rate": 3.326923076923077e-05,
|
| 1052 |
+
"loss": 2.7072,
|
| 1053 |
+
"step": 1044
|
| 1054 |
+
},
|
| 1055 |
+
{
|
| 1056 |
+
"epoch": 3.37,
|
| 1057 |
+
"learning_rate": 3.3173076923076926e-05,
|
| 1058 |
+
"loss": 2.773,
|
| 1059 |
+
"step": 1050
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 3.38,
|
| 1063 |
+
"learning_rate": 3.307692307692308e-05,
|
| 1064 |
+
"loss": 2.822,
|
| 1065 |
+
"step": 1056
|
| 1066 |
+
},
|
| 1067 |
+
{
|
| 1068 |
+
"epoch": 3.4,
|
| 1069 |
+
"learning_rate": 3.298076923076923e-05,
|
| 1070 |
+
"loss": 2.9056,
|
| 1071 |
+
"step": 1062
|
| 1072 |
+
},
|
| 1073 |
+
{
|
| 1074 |
+
"epoch": 3.42,
|
| 1075 |
+
"learning_rate": 3.288461538461539e-05,
|
| 1076 |
+
"loss": 2.749,
|
| 1077 |
+
"step": 1068
|
| 1078 |
+
},
|
| 1079 |
+
{
|
| 1080 |
+
"epoch": 3.44,
|
| 1081 |
+
"learning_rate": 3.278846153846154e-05,
|
| 1082 |
+
"loss": 2.7579,
|
| 1083 |
+
"step": 1074
|
| 1084 |
+
},
|
| 1085 |
+
{
|
| 1086 |
+
"epoch": 3.46,
|
| 1087 |
+
"learning_rate": 3.269230769230769e-05,
|
| 1088 |
+
"loss": 2.7444,
|
| 1089 |
+
"step": 1080
|
| 1090 |
+
},
|
| 1091 |
+
{
|
| 1092 |
+
"epoch": 3.48,
|
| 1093 |
+
"learning_rate": 3.2596153846153845e-05,
|
| 1094 |
+
"loss": 2.818,
|
| 1095 |
+
"step": 1086
|
| 1096 |
+
},
|
| 1097 |
+
{
|
| 1098 |
+
"epoch": 3.5,
|
| 1099 |
+
"learning_rate": 3.2500000000000004e-05,
|
| 1100 |
+
"loss": 2.6285,
|
| 1101 |
+
"step": 1092
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 3.52,
|
| 1105 |
+
"learning_rate": 3.2403846153846156e-05,
|
| 1106 |
+
"loss": 2.6473,
|
| 1107 |
+
"step": 1098
|
| 1108 |
+
},
|
| 1109 |
+
{
|
| 1110 |
+
"epoch": 3.54,
|
| 1111 |
+
"learning_rate": 3.230769230769231e-05,
|
| 1112 |
+
"loss": 2.7093,
|
| 1113 |
+
"step": 1104
|
| 1114 |
+
},
|
| 1115 |
+
{
|
| 1116 |
+
"epoch": 3.56,
|
| 1117 |
+
"learning_rate": 3.221153846153847e-05,
|
| 1118 |
+
"loss": 2.732,
|
| 1119 |
+
"step": 1110
|
| 1120 |
+
},
|
| 1121 |
+
{
|
| 1122 |
+
"epoch": 3.58,
|
| 1123 |
+
"learning_rate": 3.211538461538462e-05,
|
| 1124 |
+
"loss": 2.7064,
|
| 1125 |
+
"step": 1116
|
| 1126 |
+
},
|
| 1127 |
+
{
|
| 1128 |
+
"epoch": 3.6,
|
| 1129 |
+
"learning_rate": 3.201923076923077e-05,
|
| 1130 |
+
"loss": 2.9072,
|
| 1131 |
+
"step": 1122
|
| 1132 |
+
},
|
| 1133 |
+
{
|
| 1134 |
+
"epoch": 3.62,
|
| 1135 |
+
"learning_rate": 3.192307692307692e-05,
|
| 1136 |
+
"loss": 2.7544,
|
| 1137 |
+
"step": 1128
|
| 1138 |
+
},
|
| 1139 |
+
{
|
| 1140 |
+
"epoch": 3.63,
|
| 1141 |
+
"learning_rate": 3.1826923076923075e-05,
|
| 1142 |
+
"loss": 2.738,
|
| 1143 |
+
"step": 1134
|
| 1144 |
+
},
|
| 1145 |
+
{
|
| 1146 |
+
"epoch": 3.65,
|
| 1147 |
+
"learning_rate": 3.1730769230769234e-05,
|
| 1148 |
+
"loss": 2.6721,
|
| 1149 |
+
"step": 1140
|
| 1150 |
+
},
|
| 1151 |
+
{
|
| 1152 |
+
"epoch": 3.67,
|
| 1153 |
+
"learning_rate": 3.1634615384615386e-05,
|
| 1154 |
+
"loss": 2.8218,
|
| 1155 |
+
"step": 1146
|
| 1156 |
+
},
|
| 1157 |
+
{
|
| 1158 |
+
"epoch": 3.69,
|
| 1159 |
+
"learning_rate": 3.153846153846154e-05,
|
| 1160 |
+
"loss": 2.7553,
|
| 1161 |
+
"step": 1152
|
| 1162 |
+
},
|
| 1163 |
+
{
|
| 1164 |
+
"epoch": 3.71,
|
| 1165 |
+
"learning_rate": 3.144230769230769e-05,
|
| 1166 |
+
"loss": 2.8186,
|
| 1167 |
+
"step": 1158
|
| 1168 |
+
},
|
| 1169 |
+
{
|
| 1170 |
+
"epoch": 3.73,
|
| 1171 |
+
"learning_rate": 3.134615384615385e-05,
|
| 1172 |
+
"loss": 2.7511,
|
| 1173 |
+
"step": 1164
|
| 1174 |
+
},
|
| 1175 |
+
{
|
| 1176 |
+
"epoch": 3.75,
|
| 1177 |
+
"learning_rate": 3.125e-05,
|
| 1178 |
+
"loss": 2.6013,
|
| 1179 |
+
"step": 1170
|
| 1180 |
+
},
|
| 1181 |
+
{
|
| 1182 |
+
"epoch": 3.77,
|
| 1183 |
+
"learning_rate": 3.115384615384615e-05,
|
| 1184 |
+
"loss": 2.7191,
|
| 1185 |
+
"step": 1176
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 3.79,
|
| 1189 |
+
"learning_rate": 3.105769230769231e-05,
|
| 1190 |
+
"loss": 2.7695,
|
| 1191 |
+
"step": 1182
|
| 1192 |
+
},
|
| 1193 |
+
{
|
| 1194 |
+
"epoch": 3.81,
|
| 1195 |
+
"learning_rate": 3.0961538461538464e-05,
|
| 1196 |
+
"loss": 2.738,
|
| 1197 |
+
"step": 1188
|
| 1198 |
+
},
|
| 1199 |
+
{
|
| 1200 |
+
"epoch": 3.83,
|
| 1201 |
+
"learning_rate": 3.0865384615384616e-05,
|
| 1202 |
+
"loss": 2.7202,
|
| 1203 |
+
"step": 1194
|
| 1204 |
+
},
|
| 1205 |
+
{
|
| 1206 |
+
"epoch": 3.85,
|
| 1207 |
+
"learning_rate": 3.0769230769230774e-05,
|
| 1208 |
+
"loss": 2.6429,
|
| 1209 |
+
"step": 1200
|
| 1210 |
+
},
|
| 1211 |
+
{
|
| 1212 |
+
"epoch": 3.87,
|
| 1213 |
+
"learning_rate": 3.0673076923076926e-05,
|
| 1214 |
+
"loss": 2.6742,
|
| 1215 |
+
"step": 1206
|
| 1216 |
+
},
|
| 1217 |
+
{
|
| 1218 |
+
"epoch": 3.88,
|
| 1219 |
+
"learning_rate": 3.057692307692308e-05,
|
| 1220 |
+
"loss": 2.7048,
|
| 1221 |
+
"step": 1212
|
| 1222 |
+
},
|
| 1223 |
+
{
|
| 1224 |
+
"epoch": 3.9,
|
| 1225 |
+
"learning_rate": 3.0480769230769234e-05,
|
| 1226 |
+
"loss": 2.7016,
|
| 1227 |
+
"step": 1218
|
| 1228 |
+
},
|
| 1229 |
+
{
|
| 1230 |
+
"epoch": 3.92,
|
| 1231 |
+
"learning_rate": 3.0384615384615382e-05,
|
| 1232 |
+
"loss": 2.7375,
|
| 1233 |
+
"step": 1224
|
| 1234 |
+
},
|
| 1235 |
+
{
|
| 1236 |
+
"epoch": 3.94,
|
| 1237 |
+
"learning_rate": 3.0288461538461538e-05,
|
| 1238 |
+
"loss": 2.7926,
|
| 1239 |
+
"step": 1230
|
| 1240 |
+
},
|
| 1241 |
+
{
|
| 1242 |
+
"epoch": 3.96,
|
| 1243 |
+
"learning_rate": 3.0192307692307693e-05,
|
| 1244 |
+
"loss": 2.7038,
|
| 1245 |
+
"step": 1236
|
| 1246 |
+
},
|
| 1247 |
+
{
|
| 1248 |
+
"epoch": 3.98,
|
| 1249 |
+
"learning_rate": 3.0096153846153845e-05,
|
| 1250 |
+
"loss": 2.8294,
|
| 1251 |
+
"step": 1242
|
| 1252 |
+
},
|
| 1253 |
+
{
|
| 1254 |
+
"epoch": 4.0,
|
| 1255 |
+
"learning_rate": 3e-05,
|
| 1256 |
+
"loss": 2.7543,
|
| 1257 |
+
"step": 1248
|
| 1258 |
+
},
|
| 1259 |
+
{
|
| 1260 |
+
"epoch": 4.02,
|
| 1261 |
+
"learning_rate": 2.9903846153846156e-05,
|
| 1262 |
+
"loss": 2.6022,
|
| 1263 |
+
"step": 1254
|
| 1264 |
+
},
|
| 1265 |
+
{
|
| 1266 |
+
"epoch": 4.04,
|
| 1267 |
+
"learning_rate": 2.9807692307692308e-05,
|
| 1268 |
+
"loss": 2.6609,
|
| 1269 |
+
"step": 1260
|
| 1270 |
+
},
|
| 1271 |
+
{
|
| 1272 |
+
"epoch": 4.06,
|
| 1273 |
+
"learning_rate": 2.9711538461538464e-05,
|
| 1274 |
+
"loss": 2.6091,
|
| 1275 |
+
"step": 1266
|
| 1276 |
+
},
|
| 1277 |
+
{
|
| 1278 |
+
"epoch": 4.08,
|
| 1279 |
+
"learning_rate": 2.9615384615384616e-05,
|
| 1280 |
+
"loss": 2.5386,
|
| 1281 |
+
"step": 1272
|
| 1282 |
+
},
|
| 1283 |
+
{
|
| 1284 |
+
"epoch": 4.1,
|
| 1285 |
+
"learning_rate": 2.951923076923077e-05,
|
| 1286 |
+
"loss": 2.6655,
|
| 1287 |
+
"step": 1278
|
| 1288 |
+
},
|
| 1289 |
+
{
|
| 1290 |
+
"epoch": 4.12,
|
| 1291 |
+
"learning_rate": 2.9423076923076926e-05,
|
| 1292 |
+
"loss": 2.6023,
|
| 1293 |
+
"step": 1284
|
| 1294 |
+
},
|
| 1295 |
+
{
|
| 1296 |
+
"epoch": 4.13,
|
| 1297 |
+
"learning_rate": 2.932692307692308e-05,
|
| 1298 |
+
"loss": 2.6491,
|
| 1299 |
+
"step": 1290
|
| 1300 |
+
},
|
| 1301 |
+
{
|
| 1302 |
+
"epoch": 4.15,
|
| 1303 |
+
"learning_rate": 2.9230769230769234e-05,
|
| 1304 |
+
"loss": 2.6548,
|
| 1305 |
+
"step": 1296
|
| 1306 |
+
},
|
| 1307 |
+
{
|
| 1308 |
+
"epoch": 4.17,
|
| 1309 |
+
"learning_rate": 2.913461538461539e-05,
|
| 1310 |
+
"loss": 2.584,
|
| 1311 |
+
"step": 1302
|
| 1312 |
+
},
|
| 1313 |
+
{
|
| 1314 |
+
"epoch": 4.19,
|
| 1315 |
+
"learning_rate": 2.903846153846154e-05,
|
| 1316 |
+
"loss": 2.5447,
|
| 1317 |
+
"step": 1308
|
| 1318 |
+
},
|
| 1319 |
+
{
|
| 1320 |
+
"epoch": 4.21,
|
| 1321 |
+
"learning_rate": 2.8942307692307697e-05,
|
| 1322 |
+
"loss": 2.5931,
|
| 1323 |
+
"step": 1314
|
| 1324 |
+
},
|
| 1325 |
+
{
|
| 1326 |
+
"epoch": 4.23,
|
| 1327 |
+
"learning_rate": 2.8846153846153845e-05,
|
| 1328 |
+
"loss": 2.503,
|
| 1329 |
+
"step": 1320
|
| 1330 |
+
},
|
| 1331 |
+
{
|
| 1332 |
+
"epoch": 4.25,
|
| 1333 |
+
"learning_rate": 2.8749999999999997e-05,
|
| 1334 |
+
"loss": 2.6097,
|
| 1335 |
+
"step": 1326
|
| 1336 |
+
},
|
| 1337 |
+
{
|
| 1338 |
+
"epoch": 4.27,
|
| 1339 |
+
"learning_rate": 2.8653846153846153e-05,
|
| 1340 |
+
"loss": 2.6926,
|
| 1341 |
+
"step": 1332
|
| 1342 |
+
},
|
| 1343 |
+
{
|
| 1344 |
+
"epoch": 4.29,
|
| 1345 |
+
"learning_rate": 2.855769230769231e-05,
|
| 1346 |
+
"loss": 2.5829,
|
| 1347 |
+
"step": 1338
|
| 1348 |
+
},
|
| 1349 |
+
{
|
| 1350 |
+
"epoch": 4.31,
|
| 1351 |
+
"learning_rate": 2.846153846153846e-05,
|
| 1352 |
+
"loss": 2.6712,
|
| 1353 |
+
"step": 1344
|
| 1354 |
+
},
|
| 1355 |
+
{
|
| 1356 |
+
"epoch": 4.33,
|
| 1357 |
+
"learning_rate": 2.8365384615384616e-05,
|
| 1358 |
+
"loss": 2.3988,
|
| 1359 |
+
"step": 1350
|
| 1360 |
+
},
|
| 1361 |
+
{
|
| 1362 |
+
"epoch": 4.35,
|
| 1363 |
+
"learning_rate": 2.826923076923077e-05,
|
| 1364 |
+
"loss": 2.5714,
|
| 1365 |
+
"step": 1356
|
| 1366 |
+
},
|
| 1367 |
+
{
|
| 1368 |
+
"epoch": 4.37,
|
| 1369 |
+
"learning_rate": 2.8173076923076923e-05,
|
| 1370 |
+
"loss": 2.6627,
|
| 1371 |
+
"step": 1362
|
| 1372 |
+
},
|
| 1373 |
+
{
|
| 1374 |
+
"epoch": 4.38,
|
| 1375 |
+
"learning_rate": 2.807692307692308e-05,
|
| 1376 |
+
"loss": 2.7195,
|
| 1377 |
+
"step": 1368
|
| 1378 |
+
},
|
| 1379 |
+
{
|
| 1380 |
+
"epoch": 4.4,
|
| 1381 |
+
"learning_rate": 2.7980769230769234e-05,
|
| 1382 |
+
"loss": 2.7181,
|
| 1383 |
+
"step": 1374
|
| 1384 |
+
},
|
| 1385 |
+
{
|
| 1386 |
+
"epoch": 4.42,
|
| 1387 |
+
"learning_rate": 2.7884615384615386e-05,
|
| 1388 |
+
"loss": 2.626,
|
| 1389 |
+
"step": 1380
|
| 1390 |
+
},
|
| 1391 |
+
{
|
| 1392 |
+
"epoch": 4.44,
|
| 1393 |
+
"learning_rate": 2.778846153846154e-05,
|
| 1394 |
+
"loss": 2.638,
|
| 1395 |
+
"step": 1386
|
| 1396 |
+
},
|
| 1397 |
+
{
|
| 1398 |
+
"epoch": 4.46,
|
| 1399 |
+
"learning_rate": 2.7692307692307694e-05,
|
| 1400 |
+
"loss": 2.6275,
|
| 1401 |
+
"step": 1392
|
| 1402 |
+
},
|
| 1403 |
+
{
|
| 1404 |
+
"epoch": 4.48,
|
| 1405 |
+
"learning_rate": 2.759615384615385e-05,
|
| 1406 |
+
"loss": 2.6399,
|
| 1407 |
+
"step": 1398
|
| 1408 |
+
},
|
| 1409 |
+
{
|
| 1410 |
+
"epoch": 4.5,
|
| 1411 |
+
"learning_rate": 2.7500000000000004e-05,
|
| 1412 |
+
"loss": 2.6894,
|
| 1413 |
+
"step": 1404
|
| 1414 |
+
},
|
| 1415 |
+
{
|
| 1416 |
+
"epoch": 4.52,
|
| 1417 |
+
"learning_rate": 2.7403846153846156e-05,
|
| 1418 |
+
"loss": 2.5917,
|
| 1419 |
+
"step": 1410
|
| 1420 |
+
},
|
| 1421 |
+
{
|
| 1422 |
+
"epoch": 4.54,
|
| 1423 |
+
"learning_rate": 2.7307692307692305e-05,
|
| 1424 |
+
"loss": 2.5564,
|
| 1425 |
+
"step": 1416
|
| 1426 |
+
},
|
| 1427 |
+
{
|
| 1428 |
+
"epoch": 4.56,
|
| 1429 |
+
"learning_rate": 2.721153846153846e-05,
|
| 1430 |
+
"loss": 2.6134,
|
| 1431 |
+
"step": 1422
|
| 1432 |
+
},
|
| 1433 |
+
{
|
| 1434 |
+
"epoch": 4.58,
|
| 1435 |
+
"learning_rate": 2.7115384615384616e-05,
|
| 1436 |
+
"loss": 2.6375,
|
| 1437 |
+
"step": 1428
|
| 1438 |
+
},
|
| 1439 |
+
{
|
| 1440 |
+
"epoch": 4.6,
|
| 1441 |
+
"learning_rate": 2.7019230769230768e-05,
|
| 1442 |
+
"loss": 2.6301,
|
| 1443 |
+
"step": 1434
|
| 1444 |
+
},
|
| 1445 |
+
{
|
| 1446 |
+
"epoch": 4.62,
|
| 1447 |
+
"learning_rate": 2.6923076923076923e-05,
|
| 1448 |
+
"loss": 2.5679,
|
| 1449 |
+
"step": 1440
|
| 1450 |
+
},
|
| 1451 |
+
{
|
| 1452 |
+
"epoch": 4.63,
|
| 1453 |
+
"learning_rate": 2.682692307692308e-05,
|
| 1454 |
+
"loss": 2.6704,
|
| 1455 |
+
"step": 1446
|
| 1456 |
+
},
|
| 1457 |
+
{
|
| 1458 |
+
"epoch": 4.65,
|
| 1459 |
+
"learning_rate": 2.673076923076923e-05,
|
| 1460 |
+
"loss": 2.6778,
|
| 1461 |
+
"step": 1452
|
| 1462 |
+
},
|
| 1463 |
+
{
|
| 1464 |
+
"epoch": 4.67,
|
| 1465 |
+
"learning_rate": 2.6634615384615386e-05,
|
| 1466 |
+
"loss": 2.5626,
|
| 1467 |
+
"step": 1458
|
| 1468 |
+
},
|
| 1469 |
+
{
|
| 1470 |
+
"epoch": 4.69,
|
| 1471 |
+
"learning_rate": 2.6538461538461538e-05,
|
| 1472 |
+
"loss": 2.5557,
|
| 1473 |
+
"step": 1464
|
| 1474 |
+
},
|
| 1475 |
+
{
|
| 1476 |
+
"epoch": 4.71,
|
| 1477 |
+
"learning_rate": 2.6442307692307694e-05,
|
| 1478 |
+
"loss": 2.5245,
|
| 1479 |
+
"step": 1470
|
| 1480 |
+
},
|
| 1481 |
+
{
|
| 1482 |
+
"epoch": 4.73,
|
| 1483 |
+
"learning_rate": 2.634615384615385e-05,
|
| 1484 |
+
"loss": 2.5189,
|
| 1485 |
+
"step": 1476
|
| 1486 |
+
},
|
| 1487 |
+
{
|
| 1488 |
+
"epoch": 4.75,
|
| 1489 |
+
"learning_rate": 2.625e-05,
|
| 1490 |
+
"loss": 2.4994,
|
| 1491 |
+
"step": 1482
|
| 1492 |
+
},
|
| 1493 |
+
{
|
| 1494 |
+
"epoch": 4.77,
|
| 1495 |
+
"learning_rate": 2.6153846153846157e-05,
|
| 1496 |
+
"loss": 2.472,
|
| 1497 |
+
"step": 1488
|
| 1498 |
+
},
|
| 1499 |
+
{
|
| 1500 |
+
"epoch": 4.79,
|
| 1501 |
+
"learning_rate": 2.6057692307692312e-05,
|
| 1502 |
+
"loss": 2.5451,
|
| 1503 |
+
"step": 1494
|
| 1504 |
+
},
|
| 1505 |
+
{
|
| 1506 |
+
"epoch": 4.81,
|
| 1507 |
+
"learning_rate": 2.5961538461538464e-05,
|
| 1508 |
+
"loss": 2.5559,
|
| 1509 |
+
"step": 1500
|
| 1510 |
+
}
|
| 1511 |
+
],
|
| 1512 |
+
"logging_steps": 6,
|
| 1513 |
+
"max_steps": 3120,
|
| 1514 |
+
"num_train_epochs": 10,
|
| 1515 |
+
"save_steps": 500,
|
| 1516 |
+
"total_flos": 604654258176000.0,
|
| 1517 |
+
"trial_name": null,
|
| 1518 |
+
"trial_params": null
|
| 1519 |
+
}
|
checkpoint-1500/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c150481b2be1a59fc2f8a97e3f83518bf0fe89d39fbfc7d3d90a4bdd3f30ccaf
|
| 3 |
+
size 4600
|
checkpoint-2000/config.json
ADDED
|
@@ -0,0 +1,40 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "flax-community/gpt2-small-indonesian",
|
| 3 |
+
"activation_function": "gelu_new",
|
| 4 |
+
"architectures": [
|
| 5 |
+
"GPT2LMHeadModel"
|
| 6 |
+
],
|
| 7 |
+
"attn_pdrop": 0.0,
|
| 8 |
+
"bos_token_id": 50256,
|
| 9 |
+
"embd_pdrop": 0.0,
|
| 10 |
+
"eos_token_id": 50256,
|
| 11 |
+
"gradient_checkpointing": false,
|
| 12 |
+
"initializer_range": 0.02,
|
| 13 |
+
"layer_norm_epsilon": 1e-05,
|
| 14 |
+
"model_type": "gpt2",
|
| 15 |
+
"n_ctx": 1024,
|
| 16 |
+
"n_embd": 768,
|
| 17 |
+
"n_head": 12,
|
| 18 |
+
"n_inner": null,
|
| 19 |
+
"n_layer": 12,
|
| 20 |
+
"n_positions": 1024,
|
| 21 |
+
"reorder_and_upcast_attn": false,
|
| 22 |
+
"resid_pdrop": 0.0,
|
| 23 |
+
"scale_attn_by_inverse_layer_idx": false,
|
| 24 |
+
"scale_attn_weights": true,
|
| 25 |
+
"summary_activation": null,
|
| 26 |
+
"summary_first_dropout": 0.1,
|
| 27 |
+
"summary_proj_to_labels": true,
|
| 28 |
+
"summary_type": "cls_index",
|
| 29 |
+
"summary_use_proj": true,
|
| 30 |
+
"task_specific_params": {
|
| 31 |
+
"text-generation": {
|
| 32 |
+
"do_sample": true,
|
| 33 |
+
"max_length": 50
|
| 34 |
+
}
|
| 35 |
+
},
|
| 36 |
+
"torch_dtype": "float32",
|
| 37 |
+
"transformers_version": "4.35.2",
|
| 38 |
+
"use_cache": true,
|
| 39 |
+
"vocab_size": 50260
|
| 40 |
+
}
|
checkpoint-2000/generation_config.json
ADDED
|
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"bos_token_id": 50256,
|
| 4 |
+
"eos_token_id": 50256,
|
| 5 |
+
"transformers_version": "4.35.2"
|
| 6 |
+
}
|
checkpoint-2000/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:fee6fbbc0c5d69709e0687965fea00723388f21f73469d56064f6d69c5788976
|
| 3 |
+
size 497783424
|
checkpoint-2000/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8a23fd797a52bb1ba26e4380e9f4093086595d73d026938ca513c77bab9701ca
|
| 3 |
+
size 995660730
|
checkpoint-2000/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8d3b7102895eb0637b0cab516bd672f216b2bf79078a83eb301011a90444f44c
|
| 3 |
+
size 14244
|
checkpoint-2000/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:85e2e120a07106cab0d2ffd67d3d711ab6db96368f16f38c5df622152664e915
|
| 3 |
+
size 1064
|
checkpoint-2000/trainer_state.json
ADDED
|
@@ -0,0 +1,2017 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 6.410256410256411,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 2000,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.02,
|
| 13 |
+
"learning_rate": 4.9903846153846154e-05,
|
| 14 |
+
"loss": 5.8218,
|
| 15 |
+
"step": 6
|
| 16 |
+
},
|
| 17 |
+
{
|
| 18 |
+
"epoch": 0.04,
|
| 19 |
+
"learning_rate": 4.980769230769231e-05,
|
| 20 |
+
"loss": 5.2499,
|
| 21 |
+
"step": 12
|
| 22 |
+
},
|
| 23 |
+
{
|
| 24 |
+
"epoch": 0.06,
|
| 25 |
+
"learning_rate": 4.9711538461538465e-05,
|
| 26 |
+
"loss": 5.1236,
|
| 27 |
+
"step": 18
|
| 28 |
+
},
|
| 29 |
+
{
|
| 30 |
+
"epoch": 0.08,
|
| 31 |
+
"learning_rate": 4.961538461538462e-05,
|
| 32 |
+
"loss": 4.8363,
|
| 33 |
+
"step": 24
|
| 34 |
+
},
|
| 35 |
+
{
|
| 36 |
+
"epoch": 0.1,
|
| 37 |
+
"learning_rate": 4.9519230769230776e-05,
|
| 38 |
+
"loss": 4.5805,
|
| 39 |
+
"step": 30
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.12,
|
| 43 |
+
"learning_rate": 4.942307692307693e-05,
|
| 44 |
+
"loss": 4.5276,
|
| 45 |
+
"step": 36
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.13,
|
| 49 |
+
"learning_rate": 4.932692307692308e-05,
|
| 50 |
+
"loss": 4.3871,
|
| 51 |
+
"step": 42
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.15,
|
| 55 |
+
"learning_rate": 4.923076923076924e-05,
|
| 56 |
+
"loss": 4.2706,
|
| 57 |
+
"step": 48
|
| 58 |
+
},
|
| 59 |
+
{
|
| 60 |
+
"epoch": 0.17,
|
| 61 |
+
"learning_rate": 4.9134615384615384e-05,
|
| 62 |
+
"loss": 4.0906,
|
| 63 |
+
"step": 54
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 0.19,
|
| 67 |
+
"learning_rate": 4.9038461538461536e-05,
|
| 68 |
+
"loss": 4.1704,
|
| 69 |
+
"step": 60
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 0.21,
|
| 73 |
+
"learning_rate": 4.8942307692307695e-05,
|
| 74 |
+
"loss": 3.9014,
|
| 75 |
+
"step": 66
|
| 76 |
+
},
|
| 77 |
+
{
|
| 78 |
+
"epoch": 0.23,
|
| 79 |
+
"learning_rate": 4.884615384615385e-05,
|
| 80 |
+
"loss": 4.0338,
|
| 81 |
+
"step": 72
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"epoch": 0.25,
|
| 85 |
+
"learning_rate": 4.875e-05,
|
| 86 |
+
"loss": 4.011,
|
| 87 |
+
"step": 78
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"epoch": 0.27,
|
| 91 |
+
"learning_rate": 4.865384615384616e-05,
|
| 92 |
+
"loss": 3.8017,
|
| 93 |
+
"step": 84
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.29,
|
| 97 |
+
"learning_rate": 4.855769230769231e-05,
|
| 98 |
+
"loss": 3.819,
|
| 99 |
+
"step": 90
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"epoch": 0.31,
|
| 103 |
+
"learning_rate": 4.846153846153846e-05,
|
| 104 |
+
"loss": 3.8157,
|
| 105 |
+
"step": 96
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 0.33,
|
| 109 |
+
"learning_rate": 4.836538461538462e-05,
|
| 110 |
+
"loss": 3.7675,
|
| 111 |
+
"step": 102
|
| 112 |
+
},
|
| 113 |
+
{
|
| 114 |
+
"epoch": 0.35,
|
| 115 |
+
"learning_rate": 4.826923076923077e-05,
|
| 116 |
+
"loss": 3.756,
|
| 117 |
+
"step": 108
|
| 118 |
+
},
|
| 119 |
+
{
|
| 120 |
+
"epoch": 0.37,
|
| 121 |
+
"learning_rate": 4.8173076923076925e-05,
|
| 122 |
+
"loss": 3.7439,
|
| 123 |
+
"step": 114
|
| 124 |
+
},
|
| 125 |
+
{
|
| 126 |
+
"epoch": 0.38,
|
| 127 |
+
"learning_rate": 4.8076923076923084e-05,
|
| 128 |
+
"loss": 3.761,
|
| 129 |
+
"step": 120
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.4,
|
| 133 |
+
"learning_rate": 4.7980769230769236e-05,
|
| 134 |
+
"loss": 3.7318,
|
| 135 |
+
"step": 126
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.42,
|
| 139 |
+
"learning_rate": 4.788461538461539e-05,
|
| 140 |
+
"loss": 3.745,
|
| 141 |
+
"step": 132
|
| 142 |
+
},
|
| 143 |
+
{
|
| 144 |
+
"epoch": 0.44,
|
| 145 |
+
"learning_rate": 4.778846153846154e-05,
|
| 146 |
+
"loss": 3.784,
|
| 147 |
+
"step": 138
|
| 148 |
+
},
|
| 149 |
+
{
|
| 150 |
+
"epoch": 0.46,
|
| 151 |
+
"learning_rate": 4.76923076923077e-05,
|
| 152 |
+
"loss": 3.6114,
|
| 153 |
+
"step": 144
|
| 154 |
+
},
|
| 155 |
+
{
|
| 156 |
+
"epoch": 0.48,
|
| 157 |
+
"learning_rate": 4.7596153846153844e-05,
|
| 158 |
+
"loss": 3.637,
|
| 159 |
+
"step": 150
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 0.5,
|
| 163 |
+
"learning_rate": 4.75e-05,
|
| 164 |
+
"loss": 3.5881,
|
| 165 |
+
"step": 156
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 0.52,
|
| 169 |
+
"learning_rate": 4.7403846153846155e-05,
|
| 170 |
+
"loss": 3.5336,
|
| 171 |
+
"step": 162
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 0.54,
|
| 175 |
+
"learning_rate": 4.730769230769231e-05,
|
| 176 |
+
"loss": 3.5313,
|
| 177 |
+
"step": 168
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.56,
|
| 181 |
+
"learning_rate": 4.7211538461538465e-05,
|
| 182 |
+
"loss": 3.6357,
|
| 183 |
+
"step": 174
|
| 184 |
+
},
|
| 185 |
+
{
|
| 186 |
+
"epoch": 0.58,
|
| 187 |
+
"learning_rate": 4.711538461538462e-05,
|
| 188 |
+
"loss": 3.5818,
|
| 189 |
+
"step": 180
|
| 190 |
+
},
|
| 191 |
+
{
|
| 192 |
+
"epoch": 0.6,
|
| 193 |
+
"learning_rate": 4.701923076923077e-05,
|
| 194 |
+
"loss": 3.5017,
|
| 195 |
+
"step": 186
|
| 196 |
+
},
|
| 197 |
+
{
|
| 198 |
+
"epoch": 0.62,
|
| 199 |
+
"learning_rate": 4.692307692307693e-05,
|
| 200 |
+
"loss": 3.5815,
|
| 201 |
+
"step": 192
|
| 202 |
+
},
|
| 203 |
+
{
|
| 204 |
+
"epoch": 0.63,
|
| 205 |
+
"learning_rate": 4.682692307692308e-05,
|
| 206 |
+
"loss": 3.611,
|
| 207 |
+
"step": 198
|
| 208 |
+
},
|
| 209 |
+
{
|
| 210 |
+
"epoch": 0.65,
|
| 211 |
+
"learning_rate": 4.673076923076923e-05,
|
| 212 |
+
"loss": 3.6096,
|
| 213 |
+
"step": 204
|
| 214 |
+
},
|
| 215 |
+
{
|
| 216 |
+
"epoch": 0.67,
|
| 217 |
+
"learning_rate": 4.6634615384615384e-05,
|
| 218 |
+
"loss": 3.5436,
|
| 219 |
+
"step": 210
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.69,
|
| 223 |
+
"learning_rate": 4.653846153846154e-05,
|
| 224 |
+
"loss": 3.5258,
|
| 225 |
+
"step": 216
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 0.71,
|
| 229 |
+
"learning_rate": 4.6442307692307695e-05,
|
| 230 |
+
"loss": 3.5277,
|
| 231 |
+
"step": 222
|
| 232 |
+
},
|
| 233 |
+
{
|
| 234 |
+
"epoch": 0.73,
|
| 235 |
+
"learning_rate": 4.634615384615385e-05,
|
| 236 |
+
"loss": 3.505,
|
| 237 |
+
"step": 228
|
| 238 |
+
},
|
| 239 |
+
{
|
| 240 |
+
"epoch": 0.75,
|
| 241 |
+
"learning_rate": 4.6250000000000006e-05,
|
| 242 |
+
"loss": 3.4665,
|
| 243 |
+
"step": 234
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"epoch": 0.77,
|
| 247 |
+
"learning_rate": 4.615384615384616e-05,
|
| 248 |
+
"loss": 3.4245,
|
| 249 |
+
"step": 240
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 0.79,
|
| 253 |
+
"learning_rate": 4.605769230769231e-05,
|
| 254 |
+
"loss": 3.4547,
|
| 255 |
+
"step": 246
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 0.81,
|
| 259 |
+
"learning_rate": 4.596153846153846e-05,
|
| 260 |
+
"loss": 3.4382,
|
| 261 |
+
"step": 252
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.83,
|
| 265 |
+
"learning_rate": 4.5865384615384614e-05,
|
| 266 |
+
"loss": 3.482,
|
| 267 |
+
"step": 258
|
| 268 |
+
},
|
| 269 |
+
{
|
| 270 |
+
"epoch": 0.85,
|
| 271 |
+
"learning_rate": 4.576923076923077e-05,
|
| 272 |
+
"loss": 3.4795,
|
| 273 |
+
"step": 264
|
| 274 |
+
},
|
| 275 |
+
{
|
| 276 |
+
"epoch": 0.87,
|
| 277 |
+
"learning_rate": 4.5673076923076925e-05,
|
| 278 |
+
"loss": 3.5018,
|
| 279 |
+
"step": 270
|
| 280 |
+
},
|
| 281 |
+
{
|
| 282 |
+
"epoch": 0.88,
|
| 283 |
+
"learning_rate": 4.557692307692308e-05,
|
| 284 |
+
"loss": 3.448,
|
| 285 |
+
"step": 276
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 0.9,
|
| 289 |
+
"learning_rate": 4.548076923076923e-05,
|
| 290 |
+
"loss": 3.4206,
|
| 291 |
+
"step": 282
|
| 292 |
+
},
|
| 293 |
+
{
|
| 294 |
+
"epoch": 0.92,
|
| 295 |
+
"learning_rate": 4.538461538461539e-05,
|
| 296 |
+
"loss": 3.528,
|
| 297 |
+
"step": 288
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 0.94,
|
| 301 |
+
"learning_rate": 4.528846153846154e-05,
|
| 302 |
+
"loss": 3.475,
|
| 303 |
+
"step": 294
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.96,
|
| 307 |
+
"learning_rate": 4.519230769230769e-05,
|
| 308 |
+
"loss": 3.4142,
|
| 309 |
+
"step": 300
|
| 310 |
+
},
|
| 311 |
+
{
|
| 312 |
+
"epoch": 0.98,
|
| 313 |
+
"learning_rate": 4.509615384615385e-05,
|
| 314 |
+
"loss": 3.4156,
|
| 315 |
+
"step": 306
|
| 316 |
+
},
|
| 317 |
+
{
|
| 318 |
+
"epoch": 1.0,
|
| 319 |
+
"learning_rate": 4.5e-05,
|
| 320 |
+
"loss": 3.3274,
|
| 321 |
+
"step": 312
|
| 322 |
+
},
|
| 323 |
+
{
|
| 324 |
+
"epoch": 1.02,
|
| 325 |
+
"learning_rate": 4.4903846153846155e-05,
|
| 326 |
+
"loss": 3.3594,
|
| 327 |
+
"step": 318
|
| 328 |
+
},
|
| 329 |
+
{
|
| 330 |
+
"epoch": 1.04,
|
| 331 |
+
"learning_rate": 4.4807692307692314e-05,
|
| 332 |
+
"loss": 3.2517,
|
| 333 |
+
"step": 324
|
| 334 |
+
},
|
| 335 |
+
{
|
| 336 |
+
"epoch": 1.06,
|
| 337 |
+
"learning_rate": 4.4711538461538466e-05,
|
| 338 |
+
"loss": 3.2934,
|
| 339 |
+
"step": 330
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 1.08,
|
| 343 |
+
"learning_rate": 4.461538461538462e-05,
|
| 344 |
+
"loss": 3.2986,
|
| 345 |
+
"step": 336
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 1.1,
|
| 349 |
+
"learning_rate": 4.451923076923077e-05,
|
| 350 |
+
"loss": 3.2655,
|
| 351 |
+
"step": 342
|
| 352 |
+
},
|
| 353 |
+
{
|
| 354 |
+
"epoch": 1.12,
|
| 355 |
+
"learning_rate": 4.442307692307692e-05,
|
| 356 |
+
"loss": 3.4058,
|
| 357 |
+
"step": 348
|
| 358 |
+
},
|
| 359 |
+
{
|
| 360 |
+
"epoch": 1.13,
|
| 361 |
+
"learning_rate": 4.4326923076923074e-05,
|
| 362 |
+
"loss": 3.265,
|
| 363 |
+
"step": 354
|
| 364 |
+
},
|
| 365 |
+
{
|
| 366 |
+
"epoch": 1.15,
|
| 367 |
+
"learning_rate": 4.423076923076923e-05,
|
| 368 |
+
"loss": 3.3208,
|
| 369 |
+
"step": 360
|
| 370 |
+
},
|
| 371 |
+
{
|
| 372 |
+
"epoch": 1.17,
|
| 373 |
+
"learning_rate": 4.4134615384615385e-05,
|
| 374 |
+
"loss": 3.238,
|
| 375 |
+
"step": 366
|
| 376 |
+
},
|
| 377 |
+
{
|
| 378 |
+
"epoch": 1.19,
|
| 379 |
+
"learning_rate": 4.403846153846154e-05,
|
| 380 |
+
"loss": 3.2153,
|
| 381 |
+
"step": 372
|
| 382 |
+
},
|
| 383 |
+
{
|
| 384 |
+
"epoch": 1.21,
|
| 385 |
+
"learning_rate": 4.3942307692307695e-05,
|
| 386 |
+
"loss": 3.264,
|
| 387 |
+
"step": 378
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 1.23,
|
| 391 |
+
"learning_rate": 4.384615384615385e-05,
|
| 392 |
+
"loss": 3.2911,
|
| 393 |
+
"step": 384
|
| 394 |
+
},
|
| 395 |
+
{
|
| 396 |
+
"epoch": 1.25,
|
| 397 |
+
"learning_rate": 4.375e-05,
|
| 398 |
+
"loss": 3.3027,
|
| 399 |
+
"step": 390
|
| 400 |
+
},
|
| 401 |
+
{
|
| 402 |
+
"epoch": 1.27,
|
| 403 |
+
"learning_rate": 4.365384615384616e-05,
|
| 404 |
+
"loss": 3.2589,
|
| 405 |
+
"step": 396
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 1.29,
|
| 409 |
+
"learning_rate": 4.355769230769231e-05,
|
| 410 |
+
"loss": 3.3683,
|
| 411 |
+
"step": 402
|
| 412 |
+
},
|
| 413 |
+
{
|
| 414 |
+
"epoch": 1.31,
|
| 415 |
+
"learning_rate": 4.346153846153846e-05,
|
| 416 |
+
"loss": 3.2849,
|
| 417 |
+
"step": 408
|
| 418 |
+
},
|
| 419 |
+
{
|
| 420 |
+
"epoch": 1.33,
|
| 421 |
+
"learning_rate": 4.336538461538462e-05,
|
| 422 |
+
"loss": 3.2397,
|
| 423 |
+
"step": 414
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 1.35,
|
| 427 |
+
"learning_rate": 4.326923076923077e-05,
|
| 428 |
+
"loss": 3.2128,
|
| 429 |
+
"step": 420
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 1.37,
|
| 433 |
+
"learning_rate": 4.3173076923076925e-05,
|
| 434 |
+
"loss": 3.1944,
|
| 435 |
+
"step": 426
|
| 436 |
+
},
|
| 437 |
+
{
|
| 438 |
+
"epoch": 1.38,
|
| 439 |
+
"learning_rate": 4.3076923076923084e-05,
|
| 440 |
+
"loss": 3.1837,
|
| 441 |
+
"step": 432
|
| 442 |
+
},
|
| 443 |
+
{
|
| 444 |
+
"epoch": 1.4,
|
| 445 |
+
"learning_rate": 4.2980769230769236e-05,
|
| 446 |
+
"loss": 3.1793,
|
| 447 |
+
"step": 438
|
| 448 |
+
},
|
| 449 |
+
{
|
| 450 |
+
"epoch": 1.42,
|
| 451 |
+
"learning_rate": 4.288461538461538e-05,
|
| 452 |
+
"loss": 3.1447,
|
| 453 |
+
"step": 444
|
| 454 |
+
},
|
| 455 |
+
{
|
| 456 |
+
"epoch": 1.44,
|
| 457 |
+
"learning_rate": 4.278846153846154e-05,
|
| 458 |
+
"loss": 3.1028,
|
| 459 |
+
"step": 450
|
| 460 |
+
},
|
| 461 |
+
{
|
| 462 |
+
"epoch": 1.46,
|
| 463 |
+
"learning_rate": 4.269230769230769e-05,
|
| 464 |
+
"loss": 3.2471,
|
| 465 |
+
"step": 456
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 1.48,
|
| 469 |
+
"learning_rate": 4.2596153846153844e-05,
|
| 470 |
+
"loss": 3.1855,
|
| 471 |
+
"step": 462
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 1.5,
|
| 475 |
+
"learning_rate": 4.25e-05,
|
| 476 |
+
"loss": 3.1817,
|
| 477 |
+
"step": 468
|
| 478 |
+
},
|
| 479 |
+
{
|
| 480 |
+
"epoch": 1.52,
|
| 481 |
+
"learning_rate": 4.2403846153846155e-05,
|
| 482 |
+
"loss": 3.214,
|
| 483 |
+
"step": 474
|
| 484 |
+
},
|
| 485 |
+
{
|
| 486 |
+
"epoch": 1.54,
|
| 487 |
+
"learning_rate": 4.230769230769231e-05,
|
| 488 |
+
"loss": 2.997,
|
| 489 |
+
"step": 480
|
| 490 |
+
},
|
| 491 |
+
{
|
| 492 |
+
"epoch": 1.56,
|
| 493 |
+
"learning_rate": 4.2211538461538466e-05,
|
| 494 |
+
"loss": 3.2059,
|
| 495 |
+
"step": 486
|
| 496 |
+
},
|
| 497 |
+
{
|
| 498 |
+
"epoch": 1.58,
|
| 499 |
+
"learning_rate": 4.211538461538462e-05,
|
| 500 |
+
"loss": 3.1517,
|
| 501 |
+
"step": 492
|
| 502 |
+
},
|
| 503 |
+
{
|
| 504 |
+
"epoch": 1.6,
|
| 505 |
+
"learning_rate": 4.201923076923077e-05,
|
| 506 |
+
"loss": 3.2236,
|
| 507 |
+
"step": 498
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 1.62,
|
| 511 |
+
"learning_rate": 4.192307692307693e-05,
|
| 512 |
+
"loss": 3.1939,
|
| 513 |
+
"step": 504
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 1.63,
|
| 517 |
+
"learning_rate": 4.182692307692308e-05,
|
| 518 |
+
"loss": 3.1143,
|
| 519 |
+
"step": 510
|
| 520 |
+
},
|
| 521 |
+
{
|
| 522 |
+
"epoch": 1.65,
|
| 523 |
+
"learning_rate": 4.173076923076923e-05,
|
| 524 |
+
"loss": 3.2261,
|
| 525 |
+
"step": 516
|
| 526 |
+
},
|
| 527 |
+
{
|
| 528 |
+
"epoch": 1.67,
|
| 529 |
+
"learning_rate": 4.163461538461539e-05,
|
| 530 |
+
"loss": 3.2095,
|
| 531 |
+
"step": 522
|
| 532 |
+
},
|
| 533 |
+
{
|
| 534 |
+
"epoch": 1.69,
|
| 535 |
+
"learning_rate": 4.1538461538461544e-05,
|
| 536 |
+
"loss": 3.2498,
|
| 537 |
+
"step": 528
|
| 538 |
+
},
|
| 539 |
+
{
|
| 540 |
+
"epoch": 1.71,
|
| 541 |
+
"learning_rate": 4.1442307692307696e-05,
|
| 542 |
+
"loss": 3.1493,
|
| 543 |
+
"step": 534
|
| 544 |
+
},
|
| 545 |
+
{
|
| 546 |
+
"epoch": 1.73,
|
| 547 |
+
"learning_rate": 4.134615384615385e-05,
|
| 548 |
+
"loss": 3.0919,
|
| 549 |
+
"step": 540
|
| 550 |
+
},
|
| 551 |
+
{
|
| 552 |
+
"epoch": 1.75,
|
| 553 |
+
"learning_rate": 4.125e-05,
|
| 554 |
+
"loss": 3.2617,
|
| 555 |
+
"step": 546
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 1.77,
|
| 559 |
+
"learning_rate": 4.115384615384615e-05,
|
| 560 |
+
"loss": 3.1447,
|
| 561 |
+
"step": 552
|
| 562 |
+
},
|
| 563 |
+
{
|
| 564 |
+
"epoch": 1.79,
|
| 565 |
+
"learning_rate": 4.105769230769231e-05,
|
| 566 |
+
"loss": 3.2025,
|
| 567 |
+
"step": 558
|
| 568 |
+
},
|
| 569 |
+
{
|
| 570 |
+
"epoch": 1.81,
|
| 571 |
+
"learning_rate": 4.096153846153846e-05,
|
| 572 |
+
"loss": 3.1829,
|
| 573 |
+
"step": 564
|
| 574 |
+
},
|
| 575 |
+
{
|
| 576 |
+
"epoch": 1.83,
|
| 577 |
+
"learning_rate": 4.0865384615384615e-05,
|
| 578 |
+
"loss": 3.2216,
|
| 579 |
+
"step": 570
|
| 580 |
+
},
|
| 581 |
+
{
|
| 582 |
+
"epoch": 1.85,
|
| 583 |
+
"learning_rate": 4.0769230769230773e-05,
|
| 584 |
+
"loss": 3.1391,
|
| 585 |
+
"step": 576
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 1.87,
|
| 589 |
+
"learning_rate": 4.0673076923076926e-05,
|
| 590 |
+
"loss": 3.1344,
|
| 591 |
+
"step": 582
|
| 592 |
+
},
|
| 593 |
+
{
|
| 594 |
+
"epoch": 1.88,
|
| 595 |
+
"learning_rate": 4.057692307692308e-05,
|
| 596 |
+
"loss": 3.1699,
|
| 597 |
+
"step": 588
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 1.9,
|
| 601 |
+
"learning_rate": 4.0480769230769236e-05,
|
| 602 |
+
"loss": 3.0138,
|
| 603 |
+
"step": 594
|
| 604 |
+
},
|
| 605 |
+
{
|
| 606 |
+
"epoch": 1.92,
|
| 607 |
+
"learning_rate": 4.038461538461539e-05,
|
| 608 |
+
"loss": 3.1157,
|
| 609 |
+
"step": 600
|
| 610 |
+
},
|
| 611 |
+
{
|
| 612 |
+
"epoch": 1.94,
|
| 613 |
+
"learning_rate": 4.028846153846154e-05,
|
| 614 |
+
"loss": 3.1173,
|
| 615 |
+
"step": 606
|
| 616 |
+
},
|
| 617 |
+
{
|
| 618 |
+
"epoch": 1.96,
|
| 619 |
+
"learning_rate": 4.019230769230769e-05,
|
| 620 |
+
"loss": 3.2019,
|
| 621 |
+
"step": 612
|
| 622 |
+
},
|
| 623 |
+
{
|
| 624 |
+
"epoch": 1.98,
|
| 625 |
+
"learning_rate": 4.009615384615385e-05,
|
| 626 |
+
"loss": 3.0801,
|
| 627 |
+
"step": 618
|
| 628 |
+
},
|
| 629 |
+
{
|
| 630 |
+
"epoch": 2.0,
|
| 631 |
+
"learning_rate": 4e-05,
|
| 632 |
+
"loss": 3.0888,
|
| 633 |
+
"step": 624
|
| 634 |
+
},
|
| 635 |
+
{
|
| 636 |
+
"epoch": 2.02,
|
| 637 |
+
"learning_rate": 3.9903846153846155e-05,
|
| 638 |
+
"loss": 3.0332,
|
| 639 |
+
"step": 630
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 2.04,
|
| 643 |
+
"learning_rate": 3.980769230769231e-05,
|
| 644 |
+
"loss": 2.9418,
|
| 645 |
+
"step": 636
|
| 646 |
+
},
|
| 647 |
+
{
|
| 648 |
+
"epoch": 2.06,
|
| 649 |
+
"learning_rate": 3.971153846153846e-05,
|
| 650 |
+
"loss": 2.933,
|
| 651 |
+
"step": 642
|
| 652 |
+
},
|
| 653 |
+
{
|
| 654 |
+
"epoch": 2.08,
|
| 655 |
+
"learning_rate": 3.961538461538462e-05,
|
| 656 |
+
"loss": 3.0467,
|
| 657 |
+
"step": 648
|
| 658 |
+
},
|
| 659 |
+
{
|
| 660 |
+
"epoch": 2.1,
|
| 661 |
+
"learning_rate": 3.951923076923077e-05,
|
| 662 |
+
"loss": 3.0991,
|
| 663 |
+
"step": 654
|
| 664 |
+
},
|
| 665 |
+
{
|
| 666 |
+
"epoch": 2.12,
|
| 667 |
+
"learning_rate": 3.942307692307692e-05,
|
| 668 |
+
"loss": 2.913,
|
| 669 |
+
"step": 660
|
| 670 |
+
},
|
| 671 |
+
{
|
| 672 |
+
"epoch": 2.13,
|
| 673 |
+
"learning_rate": 3.932692307692308e-05,
|
| 674 |
+
"loss": 3.0531,
|
| 675 |
+
"step": 666
|
| 676 |
+
},
|
| 677 |
+
{
|
| 678 |
+
"epoch": 2.15,
|
| 679 |
+
"learning_rate": 3.923076923076923e-05,
|
| 680 |
+
"loss": 2.9838,
|
| 681 |
+
"step": 672
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 2.17,
|
| 685 |
+
"learning_rate": 3.9134615384615385e-05,
|
| 686 |
+
"loss": 2.9406,
|
| 687 |
+
"step": 678
|
| 688 |
+
},
|
| 689 |
+
{
|
| 690 |
+
"epoch": 2.19,
|
| 691 |
+
"learning_rate": 3.903846153846154e-05,
|
| 692 |
+
"loss": 3.046,
|
| 693 |
+
"step": 684
|
| 694 |
+
},
|
| 695 |
+
{
|
| 696 |
+
"epoch": 2.21,
|
| 697 |
+
"learning_rate": 3.8942307692307696e-05,
|
| 698 |
+
"loss": 3.006,
|
| 699 |
+
"step": 690
|
| 700 |
+
},
|
| 701 |
+
{
|
| 702 |
+
"epoch": 2.23,
|
| 703 |
+
"learning_rate": 3.884615384615385e-05,
|
| 704 |
+
"loss": 2.9774,
|
| 705 |
+
"step": 696
|
| 706 |
+
},
|
| 707 |
+
{
|
| 708 |
+
"epoch": 2.25,
|
| 709 |
+
"learning_rate": 3.875e-05,
|
| 710 |
+
"loss": 2.9937,
|
| 711 |
+
"step": 702
|
| 712 |
+
},
|
| 713 |
+
{
|
| 714 |
+
"epoch": 2.27,
|
| 715 |
+
"learning_rate": 3.865384615384616e-05,
|
| 716 |
+
"loss": 2.9737,
|
| 717 |
+
"step": 708
|
| 718 |
+
},
|
| 719 |
+
{
|
| 720 |
+
"epoch": 2.29,
|
| 721 |
+
"learning_rate": 3.855769230769231e-05,
|
| 722 |
+
"loss": 2.9772,
|
| 723 |
+
"step": 714
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 2.31,
|
| 727 |
+
"learning_rate": 3.846153846153846e-05,
|
| 728 |
+
"loss": 2.8405,
|
| 729 |
+
"step": 720
|
| 730 |
+
},
|
| 731 |
+
{
|
| 732 |
+
"epoch": 2.33,
|
| 733 |
+
"learning_rate": 3.836538461538462e-05,
|
| 734 |
+
"loss": 3.0856,
|
| 735 |
+
"step": 726
|
| 736 |
+
},
|
| 737 |
+
{
|
| 738 |
+
"epoch": 2.35,
|
| 739 |
+
"learning_rate": 3.826923076923077e-05,
|
| 740 |
+
"loss": 2.9947,
|
| 741 |
+
"step": 732
|
| 742 |
+
},
|
| 743 |
+
{
|
| 744 |
+
"epoch": 2.37,
|
| 745 |
+
"learning_rate": 3.8173076923076926e-05,
|
| 746 |
+
"loss": 2.9251,
|
| 747 |
+
"step": 738
|
| 748 |
+
},
|
| 749 |
+
{
|
| 750 |
+
"epoch": 2.38,
|
| 751 |
+
"learning_rate": 3.807692307692308e-05,
|
| 752 |
+
"loss": 2.9613,
|
| 753 |
+
"step": 744
|
| 754 |
+
},
|
| 755 |
+
{
|
| 756 |
+
"epoch": 2.4,
|
| 757 |
+
"learning_rate": 3.798076923076923e-05,
|
| 758 |
+
"loss": 2.9402,
|
| 759 |
+
"step": 750
|
| 760 |
+
},
|
| 761 |
+
{
|
| 762 |
+
"epoch": 2.42,
|
| 763 |
+
"learning_rate": 3.788461538461538e-05,
|
| 764 |
+
"loss": 2.8823,
|
| 765 |
+
"step": 756
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 2.44,
|
| 769 |
+
"learning_rate": 3.778846153846154e-05,
|
| 770 |
+
"loss": 2.8906,
|
| 771 |
+
"step": 762
|
| 772 |
+
},
|
| 773 |
+
{
|
| 774 |
+
"epoch": 2.46,
|
| 775 |
+
"learning_rate": 3.769230769230769e-05,
|
| 776 |
+
"loss": 2.9475,
|
| 777 |
+
"step": 768
|
| 778 |
+
},
|
| 779 |
+
{
|
| 780 |
+
"epoch": 2.48,
|
| 781 |
+
"learning_rate": 3.7596153846153845e-05,
|
| 782 |
+
"loss": 2.9999,
|
| 783 |
+
"step": 774
|
| 784 |
+
},
|
| 785 |
+
{
|
| 786 |
+
"epoch": 2.5,
|
| 787 |
+
"learning_rate": 3.7500000000000003e-05,
|
| 788 |
+
"loss": 2.9503,
|
| 789 |
+
"step": 780
|
| 790 |
+
},
|
| 791 |
+
{
|
| 792 |
+
"epoch": 2.52,
|
| 793 |
+
"learning_rate": 3.7403846153846156e-05,
|
| 794 |
+
"loss": 3.0112,
|
| 795 |
+
"step": 786
|
| 796 |
+
},
|
| 797 |
+
{
|
| 798 |
+
"epoch": 2.54,
|
| 799 |
+
"learning_rate": 3.730769230769231e-05,
|
| 800 |
+
"loss": 3.0096,
|
| 801 |
+
"step": 792
|
| 802 |
+
},
|
| 803 |
+
{
|
| 804 |
+
"epoch": 2.56,
|
| 805 |
+
"learning_rate": 3.7211538461538466e-05,
|
| 806 |
+
"loss": 2.9804,
|
| 807 |
+
"step": 798
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 2.58,
|
| 811 |
+
"learning_rate": 3.711538461538462e-05,
|
| 812 |
+
"loss": 2.9203,
|
| 813 |
+
"step": 804
|
| 814 |
+
},
|
| 815 |
+
{
|
| 816 |
+
"epoch": 2.6,
|
| 817 |
+
"learning_rate": 3.701923076923077e-05,
|
| 818 |
+
"loss": 2.939,
|
| 819 |
+
"step": 810
|
| 820 |
+
},
|
| 821 |
+
{
|
| 822 |
+
"epoch": 2.62,
|
| 823 |
+
"learning_rate": 3.692307692307693e-05,
|
| 824 |
+
"loss": 2.8455,
|
| 825 |
+
"step": 816
|
| 826 |
+
},
|
| 827 |
+
{
|
| 828 |
+
"epoch": 2.63,
|
| 829 |
+
"learning_rate": 3.682692307692308e-05,
|
| 830 |
+
"loss": 2.9651,
|
| 831 |
+
"step": 822
|
| 832 |
+
},
|
| 833 |
+
{
|
| 834 |
+
"epoch": 2.65,
|
| 835 |
+
"learning_rate": 3.673076923076923e-05,
|
| 836 |
+
"loss": 2.9528,
|
| 837 |
+
"step": 828
|
| 838 |
+
},
|
| 839 |
+
{
|
| 840 |
+
"epoch": 2.67,
|
| 841 |
+
"learning_rate": 3.6634615384615385e-05,
|
| 842 |
+
"loss": 2.8042,
|
| 843 |
+
"step": 834
|
| 844 |
+
},
|
| 845 |
+
{
|
| 846 |
+
"epoch": 2.69,
|
| 847 |
+
"learning_rate": 3.653846153846154e-05,
|
| 848 |
+
"loss": 2.8311,
|
| 849 |
+
"step": 840
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 2.71,
|
| 853 |
+
"learning_rate": 3.644230769230769e-05,
|
| 854 |
+
"loss": 2.8888,
|
| 855 |
+
"step": 846
|
| 856 |
+
},
|
| 857 |
+
{
|
| 858 |
+
"epoch": 2.73,
|
| 859 |
+
"learning_rate": 3.634615384615385e-05,
|
| 860 |
+
"loss": 2.9151,
|
| 861 |
+
"step": 852
|
| 862 |
+
},
|
| 863 |
+
{
|
| 864 |
+
"epoch": 2.75,
|
| 865 |
+
"learning_rate": 3.625e-05,
|
| 866 |
+
"loss": 2.9463,
|
| 867 |
+
"step": 858
|
| 868 |
+
},
|
| 869 |
+
{
|
| 870 |
+
"epoch": 2.77,
|
| 871 |
+
"learning_rate": 3.615384615384615e-05,
|
| 872 |
+
"loss": 2.957,
|
| 873 |
+
"step": 864
|
| 874 |
+
},
|
| 875 |
+
{
|
| 876 |
+
"epoch": 2.79,
|
| 877 |
+
"learning_rate": 3.605769230769231e-05,
|
| 878 |
+
"loss": 2.9473,
|
| 879 |
+
"step": 870
|
| 880 |
+
},
|
| 881 |
+
{
|
| 882 |
+
"epoch": 2.81,
|
| 883 |
+
"learning_rate": 3.596153846153846e-05,
|
| 884 |
+
"loss": 2.9994,
|
| 885 |
+
"step": 876
|
| 886 |
+
},
|
| 887 |
+
{
|
| 888 |
+
"epoch": 2.83,
|
| 889 |
+
"learning_rate": 3.5865384615384615e-05,
|
| 890 |
+
"loss": 3.0486,
|
| 891 |
+
"step": 882
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 2.85,
|
| 895 |
+
"learning_rate": 3.5769230769230774e-05,
|
| 896 |
+
"loss": 2.9487,
|
| 897 |
+
"step": 888
|
| 898 |
+
},
|
| 899 |
+
{
|
| 900 |
+
"epoch": 2.87,
|
| 901 |
+
"learning_rate": 3.5673076923076926e-05,
|
| 902 |
+
"loss": 3.0173,
|
| 903 |
+
"step": 894
|
| 904 |
+
},
|
| 905 |
+
{
|
| 906 |
+
"epoch": 2.88,
|
| 907 |
+
"learning_rate": 3.557692307692308e-05,
|
| 908 |
+
"loss": 2.8656,
|
| 909 |
+
"step": 900
|
| 910 |
+
},
|
| 911 |
+
{
|
| 912 |
+
"epoch": 2.9,
|
| 913 |
+
"learning_rate": 3.548076923076924e-05,
|
| 914 |
+
"loss": 2.8834,
|
| 915 |
+
"step": 906
|
| 916 |
+
},
|
| 917 |
+
{
|
| 918 |
+
"epoch": 2.92,
|
| 919 |
+
"learning_rate": 3.538461538461539e-05,
|
| 920 |
+
"loss": 2.9829,
|
| 921 |
+
"step": 912
|
| 922 |
+
},
|
| 923 |
+
{
|
| 924 |
+
"epoch": 2.94,
|
| 925 |
+
"learning_rate": 3.528846153846154e-05,
|
| 926 |
+
"loss": 3.001,
|
| 927 |
+
"step": 918
|
| 928 |
+
},
|
| 929 |
+
{
|
| 930 |
+
"epoch": 2.96,
|
| 931 |
+
"learning_rate": 3.51923076923077e-05,
|
| 932 |
+
"loss": 2.9618,
|
| 933 |
+
"step": 924
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 2.98,
|
| 937 |
+
"learning_rate": 3.5096153846153845e-05,
|
| 938 |
+
"loss": 2.7964,
|
| 939 |
+
"step": 930
|
| 940 |
+
},
|
| 941 |
+
{
|
| 942 |
+
"epoch": 3.0,
|
| 943 |
+
"learning_rate": 3.5e-05,
|
| 944 |
+
"loss": 2.89,
|
| 945 |
+
"step": 936
|
| 946 |
+
},
|
| 947 |
+
{
|
| 948 |
+
"epoch": 3.02,
|
| 949 |
+
"learning_rate": 3.4903846153846156e-05,
|
| 950 |
+
"loss": 2.7191,
|
| 951 |
+
"step": 942
|
| 952 |
+
},
|
| 953 |
+
{
|
| 954 |
+
"epoch": 3.04,
|
| 955 |
+
"learning_rate": 3.480769230769231e-05,
|
| 956 |
+
"loss": 2.7875,
|
| 957 |
+
"step": 948
|
| 958 |
+
},
|
| 959 |
+
{
|
| 960 |
+
"epoch": 3.06,
|
| 961 |
+
"learning_rate": 3.471153846153846e-05,
|
| 962 |
+
"loss": 2.8498,
|
| 963 |
+
"step": 954
|
| 964 |
+
},
|
| 965 |
+
{
|
| 966 |
+
"epoch": 3.08,
|
| 967 |
+
"learning_rate": 3.461538461538462e-05,
|
| 968 |
+
"loss": 2.7858,
|
| 969 |
+
"step": 960
|
| 970 |
+
},
|
| 971 |
+
{
|
| 972 |
+
"epoch": 3.1,
|
| 973 |
+
"learning_rate": 3.451923076923077e-05,
|
| 974 |
+
"loss": 2.7435,
|
| 975 |
+
"step": 966
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 3.12,
|
| 979 |
+
"learning_rate": 3.442307692307692e-05,
|
| 980 |
+
"loss": 2.8449,
|
| 981 |
+
"step": 972
|
| 982 |
+
},
|
| 983 |
+
{
|
| 984 |
+
"epoch": 3.13,
|
| 985 |
+
"learning_rate": 3.432692307692308e-05,
|
| 986 |
+
"loss": 2.7817,
|
| 987 |
+
"step": 978
|
| 988 |
+
},
|
| 989 |
+
{
|
| 990 |
+
"epoch": 3.15,
|
| 991 |
+
"learning_rate": 3.4230769230769234e-05,
|
| 992 |
+
"loss": 2.8797,
|
| 993 |
+
"step": 984
|
| 994 |
+
},
|
| 995 |
+
{
|
| 996 |
+
"epoch": 3.17,
|
| 997 |
+
"learning_rate": 3.4134615384615386e-05,
|
| 998 |
+
"loss": 2.7719,
|
| 999 |
+
"step": 990
|
| 1000 |
+
},
|
| 1001 |
+
{
|
| 1002 |
+
"epoch": 3.19,
|
| 1003 |
+
"learning_rate": 3.4038461538461544e-05,
|
| 1004 |
+
"loss": 2.7919,
|
| 1005 |
+
"step": 996
|
| 1006 |
+
},
|
| 1007 |
+
{
|
| 1008 |
+
"epoch": 3.21,
|
| 1009 |
+
"learning_rate": 3.3942307692307696e-05,
|
| 1010 |
+
"loss": 2.8072,
|
| 1011 |
+
"step": 1002
|
| 1012 |
+
},
|
| 1013 |
+
{
|
| 1014 |
+
"epoch": 3.23,
|
| 1015 |
+
"learning_rate": 3.384615384615385e-05,
|
| 1016 |
+
"loss": 2.7281,
|
| 1017 |
+
"step": 1008
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 3.25,
|
| 1021 |
+
"learning_rate": 3.375000000000001e-05,
|
| 1022 |
+
"loss": 2.8431,
|
| 1023 |
+
"step": 1014
|
| 1024 |
+
},
|
| 1025 |
+
{
|
| 1026 |
+
"epoch": 3.27,
|
| 1027 |
+
"learning_rate": 3.365384615384616e-05,
|
| 1028 |
+
"loss": 2.7154,
|
| 1029 |
+
"step": 1020
|
| 1030 |
+
},
|
| 1031 |
+
{
|
| 1032 |
+
"epoch": 3.29,
|
| 1033 |
+
"learning_rate": 3.3557692307692304e-05,
|
| 1034 |
+
"loss": 2.83,
|
| 1035 |
+
"step": 1026
|
| 1036 |
+
},
|
| 1037 |
+
{
|
| 1038 |
+
"epoch": 3.31,
|
| 1039 |
+
"learning_rate": 3.346153846153846e-05,
|
| 1040 |
+
"loss": 2.8434,
|
| 1041 |
+
"step": 1032
|
| 1042 |
+
},
|
| 1043 |
+
{
|
| 1044 |
+
"epoch": 3.33,
|
| 1045 |
+
"learning_rate": 3.3365384615384615e-05,
|
| 1046 |
+
"loss": 2.8399,
|
| 1047 |
+
"step": 1038
|
| 1048 |
+
},
|
| 1049 |
+
{
|
| 1050 |
+
"epoch": 3.35,
|
| 1051 |
+
"learning_rate": 3.326923076923077e-05,
|
| 1052 |
+
"loss": 2.7072,
|
| 1053 |
+
"step": 1044
|
| 1054 |
+
},
|
| 1055 |
+
{
|
| 1056 |
+
"epoch": 3.37,
|
| 1057 |
+
"learning_rate": 3.3173076923076926e-05,
|
| 1058 |
+
"loss": 2.773,
|
| 1059 |
+
"step": 1050
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 3.38,
|
| 1063 |
+
"learning_rate": 3.307692307692308e-05,
|
| 1064 |
+
"loss": 2.822,
|
| 1065 |
+
"step": 1056
|
| 1066 |
+
},
|
| 1067 |
+
{
|
| 1068 |
+
"epoch": 3.4,
|
| 1069 |
+
"learning_rate": 3.298076923076923e-05,
|
| 1070 |
+
"loss": 2.9056,
|
| 1071 |
+
"step": 1062
|
| 1072 |
+
},
|
| 1073 |
+
{
|
| 1074 |
+
"epoch": 3.42,
|
| 1075 |
+
"learning_rate": 3.288461538461539e-05,
|
| 1076 |
+
"loss": 2.749,
|
| 1077 |
+
"step": 1068
|
| 1078 |
+
},
|
| 1079 |
+
{
|
| 1080 |
+
"epoch": 3.44,
|
| 1081 |
+
"learning_rate": 3.278846153846154e-05,
|
| 1082 |
+
"loss": 2.7579,
|
| 1083 |
+
"step": 1074
|
| 1084 |
+
},
|
| 1085 |
+
{
|
| 1086 |
+
"epoch": 3.46,
|
| 1087 |
+
"learning_rate": 3.269230769230769e-05,
|
| 1088 |
+
"loss": 2.7444,
|
| 1089 |
+
"step": 1080
|
| 1090 |
+
},
|
| 1091 |
+
{
|
| 1092 |
+
"epoch": 3.48,
|
| 1093 |
+
"learning_rate": 3.2596153846153845e-05,
|
| 1094 |
+
"loss": 2.818,
|
| 1095 |
+
"step": 1086
|
| 1096 |
+
},
|
| 1097 |
+
{
|
| 1098 |
+
"epoch": 3.5,
|
| 1099 |
+
"learning_rate": 3.2500000000000004e-05,
|
| 1100 |
+
"loss": 2.6285,
|
| 1101 |
+
"step": 1092
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 3.52,
|
| 1105 |
+
"learning_rate": 3.2403846153846156e-05,
|
| 1106 |
+
"loss": 2.6473,
|
| 1107 |
+
"step": 1098
|
| 1108 |
+
},
|
| 1109 |
+
{
|
| 1110 |
+
"epoch": 3.54,
|
| 1111 |
+
"learning_rate": 3.230769230769231e-05,
|
| 1112 |
+
"loss": 2.7093,
|
| 1113 |
+
"step": 1104
|
| 1114 |
+
},
|
| 1115 |
+
{
|
| 1116 |
+
"epoch": 3.56,
|
| 1117 |
+
"learning_rate": 3.221153846153847e-05,
|
| 1118 |
+
"loss": 2.732,
|
| 1119 |
+
"step": 1110
|
| 1120 |
+
},
|
| 1121 |
+
{
|
| 1122 |
+
"epoch": 3.58,
|
| 1123 |
+
"learning_rate": 3.211538461538462e-05,
|
| 1124 |
+
"loss": 2.7064,
|
| 1125 |
+
"step": 1116
|
| 1126 |
+
},
|
| 1127 |
+
{
|
| 1128 |
+
"epoch": 3.6,
|
| 1129 |
+
"learning_rate": 3.201923076923077e-05,
|
| 1130 |
+
"loss": 2.9072,
|
| 1131 |
+
"step": 1122
|
| 1132 |
+
},
|
| 1133 |
+
{
|
| 1134 |
+
"epoch": 3.62,
|
| 1135 |
+
"learning_rate": 3.192307692307692e-05,
|
| 1136 |
+
"loss": 2.7544,
|
| 1137 |
+
"step": 1128
|
| 1138 |
+
},
|
| 1139 |
+
{
|
| 1140 |
+
"epoch": 3.63,
|
| 1141 |
+
"learning_rate": 3.1826923076923075e-05,
|
| 1142 |
+
"loss": 2.738,
|
| 1143 |
+
"step": 1134
|
| 1144 |
+
},
|
| 1145 |
+
{
|
| 1146 |
+
"epoch": 3.65,
|
| 1147 |
+
"learning_rate": 3.1730769230769234e-05,
|
| 1148 |
+
"loss": 2.6721,
|
| 1149 |
+
"step": 1140
|
| 1150 |
+
},
|
| 1151 |
+
{
|
| 1152 |
+
"epoch": 3.67,
|
| 1153 |
+
"learning_rate": 3.1634615384615386e-05,
|
| 1154 |
+
"loss": 2.8218,
|
| 1155 |
+
"step": 1146
|
| 1156 |
+
},
|
| 1157 |
+
{
|
| 1158 |
+
"epoch": 3.69,
|
| 1159 |
+
"learning_rate": 3.153846153846154e-05,
|
| 1160 |
+
"loss": 2.7553,
|
| 1161 |
+
"step": 1152
|
| 1162 |
+
},
|
| 1163 |
+
{
|
| 1164 |
+
"epoch": 3.71,
|
| 1165 |
+
"learning_rate": 3.144230769230769e-05,
|
| 1166 |
+
"loss": 2.8186,
|
| 1167 |
+
"step": 1158
|
| 1168 |
+
},
|
| 1169 |
+
{
|
| 1170 |
+
"epoch": 3.73,
|
| 1171 |
+
"learning_rate": 3.134615384615385e-05,
|
| 1172 |
+
"loss": 2.7511,
|
| 1173 |
+
"step": 1164
|
| 1174 |
+
},
|
| 1175 |
+
{
|
| 1176 |
+
"epoch": 3.75,
|
| 1177 |
+
"learning_rate": 3.125e-05,
|
| 1178 |
+
"loss": 2.6013,
|
| 1179 |
+
"step": 1170
|
| 1180 |
+
},
|
| 1181 |
+
{
|
| 1182 |
+
"epoch": 3.77,
|
| 1183 |
+
"learning_rate": 3.115384615384615e-05,
|
| 1184 |
+
"loss": 2.7191,
|
| 1185 |
+
"step": 1176
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 3.79,
|
| 1189 |
+
"learning_rate": 3.105769230769231e-05,
|
| 1190 |
+
"loss": 2.7695,
|
| 1191 |
+
"step": 1182
|
| 1192 |
+
},
|
| 1193 |
+
{
|
| 1194 |
+
"epoch": 3.81,
|
| 1195 |
+
"learning_rate": 3.0961538461538464e-05,
|
| 1196 |
+
"loss": 2.738,
|
| 1197 |
+
"step": 1188
|
| 1198 |
+
},
|
| 1199 |
+
{
|
| 1200 |
+
"epoch": 3.83,
|
| 1201 |
+
"learning_rate": 3.0865384615384616e-05,
|
| 1202 |
+
"loss": 2.7202,
|
| 1203 |
+
"step": 1194
|
| 1204 |
+
},
|
| 1205 |
+
{
|
| 1206 |
+
"epoch": 3.85,
|
| 1207 |
+
"learning_rate": 3.0769230769230774e-05,
|
| 1208 |
+
"loss": 2.6429,
|
| 1209 |
+
"step": 1200
|
| 1210 |
+
},
|
| 1211 |
+
{
|
| 1212 |
+
"epoch": 3.87,
|
| 1213 |
+
"learning_rate": 3.0673076923076926e-05,
|
| 1214 |
+
"loss": 2.6742,
|
| 1215 |
+
"step": 1206
|
| 1216 |
+
},
|
| 1217 |
+
{
|
| 1218 |
+
"epoch": 3.88,
|
| 1219 |
+
"learning_rate": 3.057692307692308e-05,
|
| 1220 |
+
"loss": 2.7048,
|
| 1221 |
+
"step": 1212
|
| 1222 |
+
},
|
| 1223 |
+
{
|
| 1224 |
+
"epoch": 3.9,
|
| 1225 |
+
"learning_rate": 3.0480769230769234e-05,
|
| 1226 |
+
"loss": 2.7016,
|
| 1227 |
+
"step": 1218
|
| 1228 |
+
},
|
| 1229 |
+
{
|
| 1230 |
+
"epoch": 3.92,
|
| 1231 |
+
"learning_rate": 3.0384615384615382e-05,
|
| 1232 |
+
"loss": 2.7375,
|
| 1233 |
+
"step": 1224
|
| 1234 |
+
},
|
| 1235 |
+
{
|
| 1236 |
+
"epoch": 3.94,
|
| 1237 |
+
"learning_rate": 3.0288461538461538e-05,
|
| 1238 |
+
"loss": 2.7926,
|
| 1239 |
+
"step": 1230
|
| 1240 |
+
},
|
| 1241 |
+
{
|
| 1242 |
+
"epoch": 3.96,
|
| 1243 |
+
"learning_rate": 3.0192307692307693e-05,
|
| 1244 |
+
"loss": 2.7038,
|
| 1245 |
+
"step": 1236
|
| 1246 |
+
},
|
| 1247 |
+
{
|
| 1248 |
+
"epoch": 3.98,
|
| 1249 |
+
"learning_rate": 3.0096153846153845e-05,
|
| 1250 |
+
"loss": 2.8294,
|
| 1251 |
+
"step": 1242
|
| 1252 |
+
},
|
| 1253 |
+
{
|
| 1254 |
+
"epoch": 4.0,
|
| 1255 |
+
"learning_rate": 3e-05,
|
| 1256 |
+
"loss": 2.7543,
|
| 1257 |
+
"step": 1248
|
| 1258 |
+
},
|
| 1259 |
+
{
|
| 1260 |
+
"epoch": 4.02,
|
| 1261 |
+
"learning_rate": 2.9903846153846156e-05,
|
| 1262 |
+
"loss": 2.6022,
|
| 1263 |
+
"step": 1254
|
| 1264 |
+
},
|
| 1265 |
+
{
|
| 1266 |
+
"epoch": 4.04,
|
| 1267 |
+
"learning_rate": 2.9807692307692308e-05,
|
| 1268 |
+
"loss": 2.6609,
|
| 1269 |
+
"step": 1260
|
| 1270 |
+
},
|
| 1271 |
+
{
|
| 1272 |
+
"epoch": 4.06,
|
| 1273 |
+
"learning_rate": 2.9711538461538464e-05,
|
| 1274 |
+
"loss": 2.6091,
|
| 1275 |
+
"step": 1266
|
| 1276 |
+
},
|
| 1277 |
+
{
|
| 1278 |
+
"epoch": 4.08,
|
| 1279 |
+
"learning_rate": 2.9615384615384616e-05,
|
| 1280 |
+
"loss": 2.5386,
|
| 1281 |
+
"step": 1272
|
| 1282 |
+
},
|
| 1283 |
+
{
|
| 1284 |
+
"epoch": 4.1,
|
| 1285 |
+
"learning_rate": 2.951923076923077e-05,
|
| 1286 |
+
"loss": 2.6655,
|
| 1287 |
+
"step": 1278
|
| 1288 |
+
},
|
| 1289 |
+
{
|
| 1290 |
+
"epoch": 4.12,
|
| 1291 |
+
"learning_rate": 2.9423076923076926e-05,
|
| 1292 |
+
"loss": 2.6023,
|
| 1293 |
+
"step": 1284
|
| 1294 |
+
},
|
| 1295 |
+
{
|
| 1296 |
+
"epoch": 4.13,
|
| 1297 |
+
"learning_rate": 2.932692307692308e-05,
|
| 1298 |
+
"loss": 2.6491,
|
| 1299 |
+
"step": 1290
|
| 1300 |
+
},
|
| 1301 |
+
{
|
| 1302 |
+
"epoch": 4.15,
|
| 1303 |
+
"learning_rate": 2.9230769230769234e-05,
|
| 1304 |
+
"loss": 2.6548,
|
| 1305 |
+
"step": 1296
|
| 1306 |
+
},
|
| 1307 |
+
{
|
| 1308 |
+
"epoch": 4.17,
|
| 1309 |
+
"learning_rate": 2.913461538461539e-05,
|
| 1310 |
+
"loss": 2.584,
|
| 1311 |
+
"step": 1302
|
| 1312 |
+
},
|
| 1313 |
+
{
|
| 1314 |
+
"epoch": 4.19,
|
| 1315 |
+
"learning_rate": 2.903846153846154e-05,
|
| 1316 |
+
"loss": 2.5447,
|
| 1317 |
+
"step": 1308
|
| 1318 |
+
},
|
| 1319 |
+
{
|
| 1320 |
+
"epoch": 4.21,
|
| 1321 |
+
"learning_rate": 2.8942307692307697e-05,
|
| 1322 |
+
"loss": 2.5931,
|
| 1323 |
+
"step": 1314
|
| 1324 |
+
},
|
| 1325 |
+
{
|
| 1326 |
+
"epoch": 4.23,
|
| 1327 |
+
"learning_rate": 2.8846153846153845e-05,
|
| 1328 |
+
"loss": 2.503,
|
| 1329 |
+
"step": 1320
|
| 1330 |
+
},
|
| 1331 |
+
{
|
| 1332 |
+
"epoch": 4.25,
|
| 1333 |
+
"learning_rate": 2.8749999999999997e-05,
|
| 1334 |
+
"loss": 2.6097,
|
| 1335 |
+
"step": 1326
|
| 1336 |
+
},
|
| 1337 |
+
{
|
| 1338 |
+
"epoch": 4.27,
|
| 1339 |
+
"learning_rate": 2.8653846153846153e-05,
|
| 1340 |
+
"loss": 2.6926,
|
| 1341 |
+
"step": 1332
|
| 1342 |
+
},
|
| 1343 |
+
{
|
| 1344 |
+
"epoch": 4.29,
|
| 1345 |
+
"learning_rate": 2.855769230769231e-05,
|
| 1346 |
+
"loss": 2.5829,
|
| 1347 |
+
"step": 1338
|
| 1348 |
+
},
|
| 1349 |
+
{
|
| 1350 |
+
"epoch": 4.31,
|
| 1351 |
+
"learning_rate": 2.846153846153846e-05,
|
| 1352 |
+
"loss": 2.6712,
|
| 1353 |
+
"step": 1344
|
| 1354 |
+
},
|
| 1355 |
+
{
|
| 1356 |
+
"epoch": 4.33,
|
| 1357 |
+
"learning_rate": 2.8365384615384616e-05,
|
| 1358 |
+
"loss": 2.3988,
|
| 1359 |
+
"step": 1350
|
| 1360 |
+
},
|
| 1361 |
+
{
|
| 1362 |
+
"epoch": 4.35,
|
| 1363 |
+
"learning_rate": 2.826923076923077e-05,
|
| 1364 |
+
"loss": 2.5714,
|
| 1365 |
+
"step": 1356
|
| 1366 |
+
},
|
| 1367 |
+
{
|
| 1368 |
+
"epoch": 4.37,
|
| 1369 |
+
"learning_rate": 2.8173076923076923e-05,
|
| 1370 |
+
"loss": 2.6627,
|
| 1371 |
+
"step": 1362
|
| 1372 |
+
},
|
| 1373 |
+
{
|
| 1374 |
+
"epoch": 4.38,
|
| 1375 |
+
"learning_rate": 2.807692307692308e-05,
|
| 1376 |
+
"loss": 2.7195,
|
| 1377 |
+
"step": 1368
|
| 1378 |
+
},
|
| 1379 |
+
{
|
| 1380 |
+
"epoch": 4.4,
|
| 1381 |
+
"learning_rate": 2.7980769230769234e-05,
|
| 1382 |
+
"loss": 2.7181,
|
| 1383 |
+
"step": 1374
|
| 1384 |
+
},
|
| 1385 |
+
{
|
| 1386 |
+
"epoch": 4.42,
|
| 1387 |
+
"learning_rate": 2.7884615384615386e-05,
|
| 1388 |
+
"loss": 2.626,
|
| 1389 |
+
"step": 1380
|
| 1390 |
+
},
|
| 1391 |
+
{
|
| 1392 |
+
"epoch": 4.44,
|
| 1393 |
+
"learning_rate": 2.778846153846154e-05,
|
| 1394 |
+
"loss": 2.638,
|
| 1395 |
+
"step": 1386
|
| 1396 |
+
},
|
| 1397 |
+
{
|
| 1398 |
+
"epoch": 4.46,
|
| 1399 |
+
"learning_rate": 2.7692307692307694e-05,
|
| 1400 |
+
"loss": 2.6275,
|
| 1401 |
+
"step": 1392
|
| 1402 |
+
},
|
| 1403 |
+
{
|
| 1404 |
+
"epoch": 4.48,
|
| 1405 |
+
"learning_rate": 2.759615384615385e-05,
|
| 1406 |
+
"loss": 2.6399,
|
| 1407 |
+
"step": 1398
|
| 1408 |
+
},
|
| 1409 |
+
{
|
| 1410 |
+
"epoch": 4.5,
|
| 1411 |
+
"learning_rate": 2.7500000000000004e-05,
|
| 1412 |
+
"loss": 2.6894,
|
| 1413 |
+
"step": 1404
|
| 1414 |
+
},
|
| 1415 |
+
{
|
| 1416 |
+
"epoch": 4.52,
|
| 1417 |
+
"learning_rate": 2.7403846153846156e-05,
|
| 1418 |
+
"loss": 2.5917,
|
| 1419 |
+
"step": 1410
|
| 1420 |
+
},
|
| 1421 |
+
{
|
| 1422 |
+
"epoch": 4.54,
|
| 1423 |
+
"learning_rate": 2.7307692307692305e-05,
|
| 1424 |
+
"loss": 2.5564,
|
| 1425 |
+
"step": 1416
|
| 1426 |
+
},
|
| 1427 |
+
{
|
| 1428 |
+
"epoch": 4.56,
|
| 1429 |
+
"learning_rate": 2.721153846153846e-05,
|
| 1430 |
+
"loss": 2.6134,
|
| 1431 |
+
"step": 1422
|
| 1432 |
+
},
|
| 1433 |
+
{
|
| 1434 |
+
"epoch": 4.58,
|
| 1435 |
+
"learning_rate": 2.7115384615384616e-05,
|
| 1436 |
+
"loss": 2.6375,
|
| 1437 |
+
"step": 1428
|
| 1438 |
+
},
|
| 1439 |
+
{
|
| 1440 |
+
"epoch": 4.6,
|
| 1441 |
+
"learning_rate": 2.7019230769230768e-05,
|
| 1442 |
+
"loss": 2.6301,
|
| 1443 |
+
"step": 1434
|
| 1444 |
+
},
|
| 1445 |
+
{
|
| 1446 |
+
"epoch": 4.62,
|
| 1447 |
+
"learning_rate": 2.6923076923076923e-05,
|
| 1448 |
+
"loss": 2.5679,
|
| 1449 |
+
"step": 1440
|
| 1450 |
+
},
|
| 1451 |
+
{
|
| 1452 |
+
"epoch": 4.63,
|
| 1453 |
+
"learning_rate": 2.682692307692308e-05,
|
| 1454 |
+
"loss": 2.6704,
|
| 1455 |
+
"step": 1446
|
| 1456 |
+
},
|
| 1457 |
+
{
|
| 1458 |
+
"epoch": 4.65,
|
| 1459 |
+
"learning_rate": 2.673076923076923e-05,
|
| 1460 |
+
"loss": 2.6778,
|
| 1461 |
+
"step": 1452
|
| 1462 |
+
},
|
| 1463 |
+
{
|
| 1464 |
+
"epoch": 4.67,
|
| 1465 |
+
"learning_rate": 2.6634615384615386e-05,
|
| 1466 |
+
"loss": 2.5626,
|
| 1467 |
+
"step": 1458
|
| 1468 |
+
},
|
| 1469 |
+
{
|
| 1470 |
+
"epoch": 4.69,
|
| 1471 |
+
"learning_rate": 2.6538461538461538e-05,
|
| 1472 |
+
"loss": 2.5557,
|
| 1473 |
+
"step": 1464
|
| 1474 |
+
},
|
| 1475 |
+
{
|
| 1476 |
+
"epoch": 4.71,
|
| 1477 |
+
"learning_rate": 2.6442307692307694e-05,
|
| 1478 |
+
"loss": 2.5245,
|
| 1479 |
+
"step": 1470
|
| 1480 |
+
},
|
| 1481 |
+
{
|
| 1482 |
+
"epoch": 4.73,
|
| 1483 |
+
"learning_rate": 2.634615384615385e-05,
|
| 1484 |
+
"loss": 2.5189,
|
| 1485 |
+
"step": 1476
|
| 1486 |
+
},
|
| 1487 |
+
{
|
| 1488 |
+
"epoch": 4.75,
|
| 1489 |
+
"learning_rate": 2.625e-05,
|
| 1490 |
+
"loss": 2.4994,
|
| 1491 |
+
"step": 1482
|
| 1492 |
+
},
|
| 1493 |
+
{
|
| 1494 |
+
"epoch": 4.77,
|
| 1495 |
+
"learning_rate": 2.6153846153846157e-05,
|
| 1496 |
+
"loss": 2.472,
|
| 1497 |
+
"step": 1488
|
| 1498 |
+
},
|
| 1499 |
+
{
|
| 1500 |
+
"epoch": 4.79,
|
| 1501 |
+
"learning_rate": 2.6057692307692312e-05,
|
| 1502 |
+
"loss": 2.5451,
|
| 1503 |
+
"step": 1494
|
| 1504 |
+
},
|
| 1505 |
+
{
|
| 1506 |
+
"epoch": 4.81,
|
| 1507 |
+
"learning_rate": 2.5961538461538464e-05,
|
| 1508 |
+
"loss": 2.5559,
|
| 1509 |
+
"step": 1500
|
| 1510 |
+
},
|
| 1511 |
+
{
|
| 1512 |
+
"epoch": 4.83,
|
| 1513 |
+
"learning_rate": 2.586538461538462e-05,
|
| 1514 |
+
"loss": 2.5902,
|
| 1515 |
+
"step": 1506
|
| 1516 |
+
},
|
| 1517 |
+
{
|
| 1518 |
+
"epoch": 4.85,
|
| 1519 |
+
"learning_rate": 2.5769230769230768e-05,
|
| 1520 |
+
"loss": 2.6841,
|
| 1521 |
+
"step": 1512
|
| 1522 |
+
},
|
| 1523 |
+
{
|
| 1524 |
+
"epoch": 4.87,
|
| 1525 |
+
"learning_rate": 2.5673076923076923e-05,
|
| 1526 |
+
"loss": 2.6092,
|
| 1527 |
+
"step": 1518
|
| 1528 |
+
},
|
| 1529 |
+
{
|
| 1530 |
+
"epoch": 4.88,
|
| 1531 |
+
"learning_rate": 2.5576923076923075e-05,
|
| 1532 |
+
"loss": 2.5646,
|
| 1533 |
+
"step": 1524
|
| 1534 |
+
},
|
| 1535 |
+
{
|
| 1536 |
+
"epoch": 4.9,
|
| 1537 |
+
"learning_rate": 2.548076923076923e-05,
|
| 1538 |
+
"loss": 2.5201,
|
| 1539 |
+
"step": 1530
|
| 1540 |
+
},
|
| 1541 |
+
{
|
| 1542 |
+
"epoch": 4.92,
|
| 1543 |
+
"learning_rate": 2.5384615384615383e-05,
|
| 1544 |
+
"loss": 2.5622,
|
| 1545 |
+
"step": 1536
|
| 1546 |
+
},
|
| 1547 |
+
{
|
| 1548 |
+
"epoch": 4.94,
|
| 1549 |
+
"learning_rate": 2.528846153846154e-05,
|
| 1550 |
+
"loss": 2.6323,
|
| 1551 |
+
"step": 1542
|
| 1552 |
+
},
|
| 1553 |
+
{
|
| 1554 |
+
"epoch": 4.96,
|
| 1555 |
+
"learning_rate": 2.5192307692307694e-05,
|
| 1556 |
+
"loss": 2.7179,
|
| 1557 |
+
"step": 1548
|
| 1558 |
+
},
|
| 1559 |
+
{
|
| 1560 |
+
"epoch": 4.98,
|
| 1561 |
+
"learning_rate": 2.5096153846153846e-05,
|
| 1562 |
+
"loss": 2.6441,
|
| 1563 |
+
"step": 1554
|
| 1564 |
+
},
|
| 1565 |
+
{
|
| 1566 |
+
"epoch": 5.0,
|
| 1567 |
+
"learning_rate": 2.5e-05,
|
| 1568 |
+
"loss": 2.668,
|
| 1569 |
+
"step": 1560
|
| 1570 |
+
},
|
| 1571 |
+
{
|
| 1572 |
+
"epoch": 5.02,
|
| 1573 |
+
"learning_rate": 2.4903846153846157e-05,
|
| 1574 |
+
"loss": 2.5535,
|
| 1575 |
+
"step": 1566
|
| 1576 |
+
},
|
| 1577 |
+
{
|
| 1578 |
+
"epoch": 5.04,
|
| 1579 |
+
"learning_rate": 2.480769230769231e-05,
|
| 1580 |
+
"loss": 2.4569,
|
| 1581 |
+
"step": 1572
|
| 1582 |
+
},
|
| 1583 |
+
{
|
| 1584 |
+
"epoch": 5.06,
|
| 1585 |
+
"learning_rate": 2.4711538461538464e-05,
|
| 1586 |
+
"loss": 2.5378,
|
| 1587 |
+
"step": 1578
|
| 1588 |
+
},
|
| 1589 |
+
{
|
| 1590 |
+
"epoch": 5.08,
|
| 1591 |
+
"learning_rate": 2.461538461538462e-05,
|
| 1592 |
+
"loss": 2.537,
|
| 1593 |
+
"step": 1584
|
| 1594 |
+
},
|
| 1595 |
+
{
|
| 1596 |
+
"epoch": 5.1,
|
| 1597 |
+
"learning_rate": 2.4519230769230768e-05,
|
| 1598 |
+
"loss": 2.4913,
|
| 1599 |
+
"step": 1590
|
| 1600 |
+
},
|
| 1601 |
+
{
|
| 1602 |
+
"epoch": 5.12,
|
| 1603 |
+
"learning_rate": 2.4423076923076924e-05,
|
| 1604 |
+
"loss": 2.5156,
|
| 1605 |
+
"step": 1596
|
| 1606 |
+
},
|
| 1607 |
+
{
|
| 1608 |
+
"epoch": 5.13,
|
| 1609 |
+
"learning_rate": 2.432692307692308e-05,
|
| 1610 |
+
"loss": 2.493,
|
| 1611 |
+
"step": 1602
|
| 1612 |
+
},
|
| 1613 |
+
{
|
| 1614 |
+
"epoch": 5.15,
|
| 1615 |
+
"learning_rate": 2.423076923076923e-05,
|
| 1616 |
+
"loss": 2.5357,
|
| 1617 |
+
"step": 1608
|
| 1618 |
+
},
|
| 1619 |
+
{
|
| 1620 |
+
"epoch": 5.17,
|
| 1621 |
+
"learning_rate": 2.4134615384615386e-05,
|
| 1622 |
+
"loss": 2.4508,
|
| 1623 |
+
"step": 1614
|
| 1624 |
+
},
|
| 1625 |
+
{
|
| 1626 |
+
"epoch": 5.19,
|
| 1627 |
+
"learning_rate": 2.4038461538461542e-05,
|
| 1628 |
+
"loss": 2.5045,
|
| 1629 |
+
"step": 1620
|
| 1630 |
+
},
|
| 1631 |
+
{
|
| 1632 |
+
"epoch": 5.21,
|
| 1633 |
+
"learning_rate": 2.3942307692307694e-05,
|
| 1634 |
+
"loss": 2.4617,
|
| 1635 |
+
"step": 1626
|
| 1636 |
+
},
|
| 1637 |
+
{
|
| 1638 |
+
"epoch": 5.23,
|
| 1639 |
+
"learning_rate": 2.384615384615385e-05,
|
| 1640 |
+
"loss": 2.5392,
|
| 1641 |
+
"step": 1632
|
| 1642 |
+
},
|
| 1643 |
+
{
|
| 1644 |
+
"epoch": 5.25,
|
| 1645 |
+
"learning_rate": 2.375e-05,
|
| 1646 |
+
"loss": 2.3976,
|
| 1647 |
+
"step": 1638
|
| 1648 |
+
},
|
| 1649 |
+
{
|
| 1650 |
+
"epoch": 5.27,
|
| 1651 |
+
"learning_rate": 2.3653846153846153e-05,
|
| 1652 |
+
"loss": 2.5288,
|
| 1653 |
+
"step": 1644
|
| 1654 |
+
},
|
| 1655 |
+
{
|
| 1656 |
+
"epoch": 5.29,
|
| 1657 |
+
"learning_rate": 2.355769230769231e-05,
|
| 1658 |
+
"loss": 2.5243,
|
| 1659 |
+
"step": 1650
|
| 1660 |
+
},
|
| 1661 |
+
{
|
| 1662 |
+
"epoch": 5.31,
|
| 1663 |
+
"learning_rate": 2.3461538461538464e-05,
|
| 1664 |
+
"loss": 2.5018,
|
| 1665 |
+
"step": 1656
|
| 1666 |
+
},
|
| 1667 |
+
{
|
| 1668 |
+
"epoch": 5.33,
|
| 1669 |
+
"learning_rate": 2.3365384615384616e-05,
|
| 1670 |
+
"loss": 2.5947,
|
| 1671 |
+
"step": 1662
|
| 1672 |
+
},
|
| 1673 |
+
{
|
| 1674 |
+
"epoch": 5.35,
|
| 1675 |
+
"learning_rate": 2.326923076923077e-05,
|
| 1676 |
+
"loss": 2.5226,
|
| 1677 |
+
"step": 1668
|
| 1678 |
+
},
|
| 1679 |
+
{
|
| 1680 |
+
"epoch": 5.37,
|
| 1681 |
+
"learning_rate": 2.3173076923076924e-05,
|
| 1682 |
+
"loss": 2.4687,
|
| 1683 |
+
"step": 1674
|
| 1684 |
+
},
|
| 1685 |
+
{
|
| 1686 |
+
"epoch": 5.38,
|
| 1687 |
+
"learning_rate": 2.307692307692308e-05,
|
| 1688 |
+
"loss": 2.5297,
|
| 1689 |
+
"step": 1680
|
| 1690 |
+
},
|
| 1691 |
+
{
|
| 1692 |
+
"epoch": 5.4,
|
| 1693 |
+
"learning_rate": 2.298076923076923e-05,
|
| 1694 |
+
"loss": 2.5457,
|
| 1695 |
+
"step": 1686
|
| 1696 |
+
},
|
| 1697 |
+
{
|
| 1698 |
+
"epoch": 5.42,
|
| 1699 |
+
"learning_rate": 2.2884615384615387e-05,
|
| 1700 |
+
"loss": 2.4218,
|
| 1701 |
+
"step": 1692
|
| 1702 |
+
},
|
| 1703 |
+
{
|
| 1704 |
+
"epoch": 5.44,
|
| 1705 |
+
"learning_rate": 2.278846153846154e-05,
|
| 1706 |
+
"loss": 2.5219,
|
| 1707 |
+
"step": 1698
|
| 1708 |
+
},
|
| 1709 |
+
{
|
| 1710 |
+
"epoch": 5.46,
|
| 1711 |
+
"learning_rate": 2.2692307692307694e-05,
|
| 1712 |
+
"loss": 2.4858,
|
| 1713 |
+
"step": 1704
|
| 1714 |
+
},
|
| 1715 |
+
{
|
| 1716 |
+
"epoch": 5.48,
|
| 1717 |
+
"learning_rate": 2.2596153846153846e-05,
|
| 1718 |
+
"loss": 2.4715,
|
| 1719 |
+
"step": 1710
|
| 1720 |
+
},
|
| 1721 |
+
{
|
| 1722 |
+
"epoch": 5.5,
|
| 1723 |
+
"learning_rate": 2.25e-05,
|
| 1724 |
+
"loss": 2.3294,
|
| 1725 |
+
"step": 1716
|
| 1726 |
+
},
|
| 1727 |
+
{
|
| 1728 |
+
"epoch": 5.52,
|
| 1729 |
+
"learning_rate": 2.2403846153846157e-05,
|
| 1730 |
+
"loss": 2.5062,
|
| 1731 |
+
"step": 1722
|
| 1732 |
+
},
|
| 1733 |
+
{
|
| 1734 |
+
"epoch": 5.54,
|
| 1735 |
+
"learning_rate": 2.230769230769231e-05,
|
| 1736 |
+
"loss": 2.392,
|
| 1737 |
+
"step": 1728
|
| 1738 |
+
},
|
| 1739 |
+
{
|
| 1740 |
+
"epoch": 5.56,
|
| 1741 |
+
"learning_rate": 2.221153846153846e-05,
|
| 1742 |
+
"loss": 2.4092,
|
| 1743 |
+
"step": 1734
|
| 1744 |
+
},
|
| 1745 |
+
{
|
| 1746 |
+
"epoch": 5.58,
|
| 1747 |
+
"learning_rate": 2.2115384615384616e-05,
|
| 1748 |
+
"loss": 2.4257,
|
| 1749 |
+
"step": 1740
|
| 1750 |
+
},
|
| 1751 |
+
{
|
| 1752 |
+
"epoch": 5.6,
|
| 1753 |
+
"learning_rate": 2.201923076923077e-05,
|
| 1754 |
+
"loss": 2.4426,
|
| 1755 |
+
"step": 1746
|
| 1756 |
+
},
|
| 1757 |
+
{
|
| 1758 |
+
"epoch": 5.62,
|
| 1759 |
+
"learning_rate": 2.1923076923076924e-05,
|
| 1760 |
+
"loss": 2.4672,
|
| 1761 |
+
"step": 1752
|
| 1762 |
+
},
|
| 1763 |
+
{
|
| 1764 |
+
"epoch": 5.63,
|
| 1765 |
+
"learning_rate": 2.182692307692308e-05,
|
| 1766 |
+
"loss": 2.5468,
|
| 1767 |
+
"step": 1758
|
| 1768 |
+
},
|
| 1769 |
+
{
|
| 1770 |
+
"epoch": 5.65,
|
| 1771 |
+
"learning_rate": 2.173076923076923e-05,
|
| 1772 |
+
"loss": 2.3917,
|
| 1773 |
+
"step": 1764
|
| 1774 |
+
},
|
| 1775 |
+
{
|
| 1776 |
+
"epoch": 5.67,
|
| 1777 |
+
"learning_rate": 2.1634615384615387e-05,
|
| 1778 |
+
"loss": 2.3895,
|
| 1779 |
+
"step": 1770
|
| 1780 |
+
},
|
| 1781 |
+
{
|
| 1782 |
+
"epoch": 5.69,
|
| 1783 |
+
"learning_rate": 2.1538461538461542e-05,
|
| 1784 |
+
"loss": 2.5246,
|
| 1785 |
+
"step": 1776
|
| 1786 |
+
},
|
| 1787 |
+
{
|
| 1788 |
+
"epoch": 5.71,
|
| 1789 |
+
"learning_rate": 2.144230769230769e-05,
|
| 1790 |
+
"loss": 2.5025,
|
| 1791 |
+
"step": 1782
|
| 1792 |
+
},
|
| 1793 |
+
{
|
| 1794 |
+
"epoch": 5.73,
|
| 1795 |
+
"learning_rate": 2.1346153846153846e-05,
|
| 1796 |
+
"loss": 2.4898,
|
| 1797 |
+
"step": 1788
|
| 1798 |
+
},
|
| 1799 |
+
{
|
| 1800 |
+
"epoch": 5.75,
|
| 1801 |
+
"learning_rate": 2.125e-05,
|
| 1802 |
+
"loss": 2.5358,
|
| 1803 |
+
"step": 1794
|
| 1804 |
+
},
|
| 1805 |
+
{
|
| 1806 |
+
"epoch": 5.77,
|
| 1807 |
+
"learning_rate": 2.1153846153846154e-05,
|
| 1808 |
+
"loss": 2.5346,
|
| 1809 |
+
"step": 1800
|
| 1810 |
+
},
|
| 1811 |
+
{
|
| 1812 |
+
"epoch": 5.79,
|
| 1813 |
+
"learning_rate": 2.105769230769231e-05,
|
| 1814 |
+
"loss": 2.3057,
|
| 1815 |
+
"step": 1806
|
| 1816 |
+
},
|
| 1817 |
+
{
|
| 1818 |
+
"epoch": 5.81,
|
| 1819 |
+
"learning_rate": 2.0961538461538464e-05,
|
| 1820 |
+
"loss": 2.5392,
|
| 1821 |
+
"step": 1812
|
| 1822 |
+
},
|
| 1823 |
+
{
|
| 1824 |
+
"epoch": 5.83,
|
| 1825 |
+
"learning_rate": 2.0865384615384616e-05,
|
| 1826 |
+
"loss": 2.5815,
|
| 1827 |
+
"step": 1818
|
| 1828 |
+
},
|
| 1829 |
+
{
|
| 1830 |
+
"epoch": 5.85,
|
| 1831 |
+
"learning_rate": 2.0769230769230772e-05,
|
| 1832 |
+
"loss": 2.2718,
|
| 1833 |
+
"step": 1824
|
| 1834 |
+
},
|
| 1835 |
+
{
|
| 1836 |
+
"epoch": 5.87,
|
| 1837 |
+
"learning_rate": 2.0673076923076924e-05,
|
| 1838 |
+
"loss": 2.474,
|
| 1839 |
+
"step": 1830
|
| 1840 |
+
},
|
| 1841 |
+
{
|
| 1842 |
+
"epoch": 5.88,
|
| 1843 |
+
"learning_rate": 2.0576923076923076e-05,
|
| 1844 |
+
"loss": 2.4821,
|
| 1845 |
+
"step": 1836
|
| 1846 |
+
},
|
| 1847 |
+
{
|
| 1848 |
+
"epoch": 5.9,
|
| 1849 |
+
"learning_rate": 2.048076923076923e-05,
|
| 1850 |
+
"loss": 2.5334,
|
| 1851 |
+
"step": 1842
|
| 1852 |
+
},
|
| 1853 |
+
{
|
| 1854 |
+
"epoch": 5.92,
|
| 1855 |
+
"learning_rate": 2.0384615384615387e-05,
|
| 1856 |
+
"loss": 2.3877,
|
| 1857 |
+
"step": 1848
|
| 1858 |
+
},
|
| 1859 |
+
{
|
| 1860 |
+
"epoch": 5.94,
|
| 1861 |
+
"learning_rate": 2.028846153846154e-05,
|
| 1862 |
+
"loss": 2.3301,
|
| 1863 |
+
"step": 1854
|
| 1864 |
+
},
|
| 1865 |
+
{
|
| 1866 |
+
"epoch": 5.96,
|
| 1867 |
+
"learning_rate": 2.0192307692307694e-05,
|
| 1868 |
+
"loss": 2.545,
|
| 1869 |
+
"step": 1860
|
| 1870 |
+
},
|
| 1871 |
+
{
|
| 1872 |
+
"epoch": 5.98,
|
| 1873 |
+
"learning_rate": 2.0096153846153846e-05,
|
| 1874 |
+
"loss": 2.3865,
|
| 1875 |
+
"step": 1866
|
| 1876 |
+
},
|
| 1877 |
+
{
|
| 1878 |
+
"epoch": 6.0,
|
| 1879 |
+
"learning_rate": 2e-05,
|
| 1880 |
+
"loss": 2.4531,
|
| 1881 |
+
"step": 1872
|
| 1882 |
+
},
|
| 1883 |
+
{
|
| 1884 |
+
"epoch": 6.02,
|
| 1885 |
+
"learning_rate": 1.9903846153846154e-05,
|
| 1886 |
+
"loss": 2.4046,
|
| 1887 |
+
"step": 1878
|
| 1888 |
+
},
|
| 1889 |
+
{
|
| 1890 |
+
"epoch": 6.04,
|
| 1891 |
+
"learning_rate": 1.980769230769231e-05,
|
| 1892 |
+
"loss": 2.476,
|
| 1893 |
+
"step": 1884
|
| 1894 |
+
},
|
| 1895 |
+
{
|
| 1896 |
+
"epoch": 6.06,
|
| 1897 |
+
"learning_rate": 1.971153846153846e-05,
|
| 1898 |
+
"loss": 2.3547,
|
| 1899 |
+
"step": 1890
|
| 1900 |
+
},
|
| 1901 |
+
{
|
| 1902 |
+
"epoch": 6.08,
|
| 1903 |
+
"learning_rate": 1.9615384615384617e-05,
|
| 1904 |
+
"loss": 2.5396,
|
| 1905 |
+
"step": 1896
|
| 1906 |
+
},
|
| 1907 |
+
{
|
| 1908 |
+
"epoch": 6.1,
|
| 1909 |
+
"learning_rate": 1.951923076923077e-05,
|
| 1910 |
+
"loss": 2.3868,
|
| 1911 |
+
"step": 1902
|
| 1912 |
+
},
|
| 1913 |
+
{
|
| 1914 |
+
"epoch": 6.12,
|
| 1915 |
+
"learning_rate": 1.9423076923076924e-05,
|
| 1916 |
+
"loss": 2.2733,
|
| 1917 |
+
"step": 1908
|
| 1918 |
+
},
|
| 1919 |
+
{
|
| 1920 |
+
"epoch": 6.13,
|
| 1921 |
+
"learning_rate": 1.932692307692308e-05,
|
| 1922 |
+
"loss": 2.283,
|
| 1923 |
+
"step": 1914
|
| 1924 |
+
},
|
| 1925 |
+
{
|
| 1926 |
+
"epoch": 6.15,
|
| 1927 |
+
"learning_rate": 1.923076923076923e-05,
|
| 1928 |
+
"loss": 2.4231,
|
| 1929 |
+
"step": 1920
|
| 1930 |
+
},
|
| 1931 |
+
{
|
| 1932 |
+
"epoch": 6.17,
|
| 1933 |
+
"learning_rate": 1.9134615384615383e-05,
|
| 1934 |
+
"loss": 2.375,
|
| 1935 |
+
"step": 1926
|
| 1936 |
+
},
|
| 1937 |
+
{
|
| 1938 |
+
"epoch": 6.19,
|
| 1939 |
+
"learning_rate": 1.903846153846154e-05,
|
| 1940 |
+
"loss": 2.3824,
|
| 1941 |
+
"step": 1932
|
| 1942 |
+
},
|
| 1943 |
+
{
|
| 1944 |
+
"epoch": 6.21,
|
| 1945 |
+
"learning_rate": 1.894230769230769e-05,
|
| 1946 |
+
"loss": 2.3903,
|
| 1947 |
+
"step": 1938
|
| 1948 |
+
},
|
| 1949 |
+
{
|
| 1950 |
+
"epoch": 6.23,
|
| 1951 |
+
"learning_rate": 1.8846153846153846e-05,
|
| 1952 |
+
"loss": 2.2854,
|
| 1953 |
+
"step": 1944
|
| 1954 |
+
},
|
| 1955 |
+
{
|
| 1956 |
+
"epoch": 6.25,
|
| 1957 |
+
"learning_rate": 1.8750000000000002e-05,
|
| 1958 |
+
"loss": 2.2358,
|
| 1959 |
+
"step": 1950
|
| 1960 |
+
},
|
| 1961 |
+
{
|
| 1962 |
+
"epoch": 6.27,
|
| 1963 |
+
"learning_rate": 1.8653846153846154e-05,
|
| 1964 |
+
"loss": 2.4751,
|
| 1965 |
+
"step": 1956
|
| 1966 |
+
},
|
| 1967 |
+
{
|
| 1968 |
+
"epoch": 6.29,
|
| 1969 |
+
"learning_rate": 1.855769230769231e-05,
|
| 1970 |
+
"loss": 2.3434,
|
| 1971 |
+
"step": 1962
|
| 1972 |
+
},
|
| 1973 |
+
{
|
| 1974 |
+
"epoch": 6.31,
|
| 1975 |
+
"learning_rate": 1.8461538461538465e-05,
|
| 1976 |
+
"loss": 2.4088,
|
| 1977 |
+
"step": 1968
|
| 1978 |
+
},
|
| 1979 |
+
{
|
| 1980 |
+
"epoch": 6.33,
|
| 1981 |
+
"learning_rate": 1.8365384615384617e-05,
|
| 1982 |
+
"loss": 2.2746,
|
| 1983 |
+
"step": 1974
|
| 1984 |
+
},
|
| 1985 |
+
{
|
| 1986 |
+
"epoch": 6.35,
|
| 1987 |
+
"learning_rate": 1.826923076923077e-05,
|
| 1988 |
+
"loss": 2.4396,
|
| 1989 |
+
"step": 1980
|
| 1990 |
+
},
|
| 1991 |
+
{
|
| 1992 |
+
"epoch": 6.37,
|
| 1993 |
+
"learning_rate": 1.8173076923076924e-05,
|
| 1994 |
+
"loss": 2.3663,
|
| 1995 |
+
"step": 1986
|
| 1996 |
+
},
|
| 1997 |
+
{
|
| 1998 |
+
"epoch": 6.38,
|
| 1999 |
+
"learning_rate": 1.8076923076923076e-05,
|
| 2000 |
+
"loss": 2.4488,
|
| 2001 |
+
"step": 1992
|
| 2002 |
+
},
|
| 2003 |
+
{
|
| 2004 |
+
"epoch": 6.4,
|
| 2005 |
+
"learning_rate": 1.798076923076923e-05,
|
| 2006 |
+
"loss": 2.3495,
|
| 2007 |
+
"step": 1998
|
| 2008 |
+
}
|
| 2009 |
+
],
|
| 2010 |
+
"logging_steps": 6,
|
| 2011 |
+
"max_steps": 3120,
|
| 2012 |
+
"num_train_epochs": 10,
|
| 2013 |
+
"save_steps": 500,
|
| 2014 |
+
"total_flos": 805869536256000.0,
|
| 2015 |
+
"trial_name": null,
|
| 2016 |
+
"trial_params": null
|
| 2017 |
+
}
|
checkpoint-2000/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c150481b2be1a59fc2f8a97e3f83518bf0fe89d39fbfc7d3d90a4bdd3f30ccaf
|
| 3 |
+
size 4600
|
checkpoint-2500/config.json
ADDED
|
@@ -0,0 +1,40 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "flax-community/gpt2-small-indonesian",
|
| 3 |
+
"activation_function": "gelu_new",
|
| 4 |
+
"architectures": [
|
| 5 |
+
"GPT2LMHeadModel"
|
| 6 |
+
],
|
| 7 |
+
"attn_pdrop": 0.0,
|
| 8 |
+
"bos_token_id": 50256,
|
| 9 |
+
"embd_pdrop": 0.0,
|
| 10 |
+
"eos_token_id": 50256,
|
| 11 |
+
"gradient_checkpointing": false,
|
| 12 |
+
"initializer_range": 0.02,
|
| 13 |
+
"layer_norm_epsilon": 1e-05,
|
| 14 |
+
"model_type": "gpt2",
|
| 15 |
+
"n_ctx": 1024,
|
| 16 |
+
"n_embd": 768,
|
| 17 |
+
"n_head": 12,
|
| 18 |
+
"n_inner": null,
|
| 19 |
+
"n_layer": 12,
|
| 20 |
+
"n_positions": 1024,
|
| 21 |
+
"reorder_and_upcast_attn": false,
|
| 22 |
+
"resid_pdrop": 0.0,
|
| 23 |
+
"scale_attn_by_inverse_layer_idx": false,
|
| 24 |
+
"scale_attn_weights": true,
|
| 25 |
+
"summary_activation": null,
|
| 26 |
+
"summary_first_dropout": 0.1,
|
| 27 |
+
"summary_proj_to_labels": true,
|
| 28 |
+
"summary_type": "cls_index",
|
| 29 |
+
"summary_use_proj": true,
|
| 30 |
+
"task_specific_params": {
|
| 31 |
+
"text-generation": {
|
| 32 |
+
"do_sample": true,
|
| 33 |
+
"max_length": 50
|
| 34 |
+
}
|
| 35 |
+
},
|
| 36 |
+
"torch_dtype": "float32",
|
| 37 |
+
"transformers_version": "4.35.2",
|
| 38 |
+
"use_cache": true,
|
| 39 |
+
"vocab_size": 50260
|
| 40 |
+
}
|
checkpoint-2500/generation_config.json
ADDED
|
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"bos_token_id": 50256,
|
| 4 |
+
"eos_token_id": 50256,
|
| 5 |
+
"transformers_version": "4.35.2"
|
| 6 |
+
}
|
checkpoint-2500/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d2805766611f7321cf15dc2e87f60b09c1ea56ad139696d2d3ee7bcf72591443
|
| 3 |
+
size 497783424
|
checkpoint-2500/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5bec206d6d3cb4d92518138447f635d5cfdf6060a267d5f4e52c7bd39613d8c2
|
| 3 |
+
size 995660730
|
checkpoint-2500/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:adbac4581908b419c3039e502a4aacf8ebe77d7c5097a659c9e661f2ab321b78
|
| 3 |
+
size 14244
|
checkpoint-2500/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:da3c8008a520c0182566e91dab9d3bd68ed83ad8426d45abf6cfc07d29d248c3
|
| 3 |
+
size 1064
|
checkpoint-2500/trainer_state.json
ADDED
|
@@ -0,0 +1,2515 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 8.012820512820513,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 2500,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.02,
|
| 13 |
+
"learning_rate": 4.9903846153846154e-05,
|
| 14 |
+
"loss": 5.8218,
|
| 15 |
+
"step": 6
|
| 16 |
+
},
|
| 17 |
+
{
|
| 18 |
+
"epoch": 0.04,
|
| 19 |
+
"learning_rate": 4.980769230769231e-05,
|
| 20 |
+
"loss": 5.2499,
|
| 21 |
+
"step": 12
|
| 22 |
+
},
|
| 23 |
+
{
|
| 24 |
+
"epoch": 0.06,
|
| 25 |
+
"learning_rate": 4.9711538461538465e-05,
|
| 26 |
+
"loss": 5.1236,
|
| 27 |
+
"step": 18
|
| 28 |
+
},
|
| 29 |
+
{
|
| 30 |
+
"epoch": 0.08,
|
| 31 |
+
"learning_rate": 4.961538461538462e-05,
|
| 32 |
+
"loss": 4.8363,
|
| 33 |
+
"step": 24
|
| 34 |
+
},
|
| 35 |
+
{
|
| 36 |
+
"epoch": 0.1,
|
| 37 |
+
"learning_rate": 4.9519230769230776e-05,
|
| 38 |
+
"loss": 4.5805,
|
| 39 |
+
"step": 30
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.12,
|
| 43 |
+
"learning_rate": 4.942307692307693e-05,
|
| 44 |
+
"loss": 4.5276,
|
| 45 |
+
"step": 36
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.13,
|
| 49 |
+
"learning_rate": 4.932692307692308e-05,
|
| 50 |
+
"loss": 4.3871,
|
| 51 |
+
"step": 42
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.15,
|
| 55 |
+
"learning_rate": 4.923076923076924e-05,
|
| 56 |
+
"loss": 4.2706,
|
| 57 |
+
"step": 48
|
| 58 |
+
},
|
| 59 |
+
{
|
| 60 |
+
"epoch": 0.17,
|
| 61 |
+
"learning_rate": 4.9134615384615384e-05,
|
| 62 |
+
"loss": 4.0906,
|
| 63 |
+
"step": 54
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 0.19,
|
| 67 |
+
"learning_rate": 4.9038461538461536e-05,
|
| 68 |
+
"loss": 4.1704,
|
| 69 |
+
"step": 60
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 0.21,
|
| 73 |
+
"learning_rate": 4.8942307692307695e-05,
|
| 74 |
+
"loss": 3.9014,
|
| 75 |
+
"step": 66
|
| 76 |
+
},
|
| 77 |
+
{
|
| 78 |
+
"epoch": 0.23,
|
| 79 |
+
"learning_rate": 4.884615384615385e-05,
|
| 80 |
+
"loss": 4.0338,
|
| 81 |
+
"step": 72
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"epoch": 0.25,
|
| 85 |
+
"learning_rate": 4.875e-05,
|
| 86 |
+
"loss": 4.011,
|
| 87 |
+
"step": 78
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"epoch": 0.27,
|
| 91 |
+
"learning_rate": 4.865384615384616e-05,
|
| 92 |
+
"loss": 3.8017,
|
| 93 |
+
"step": 84
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.29,
|
| 97 |
+
"learning_rate": 4.855769230769231e-05,
|
| 98 |
+
"loss": 3.819,
|
| 99 |
+
"step": 90
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"epoch": 0.31,
|
| 103 |
+
"learning_rate": 4.846153846153846e-05,
|
| 104 |
+
"loss": 3.8157,
|
| 105 |
+
"step": 96
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 0.33,
|
| 109 |
+
"learning_rate": 4.836538461538462e-05,
|
| 110 |
+
"loss": 3.7675,
|
| 111 |
+
"step": 102
|
| 112 |
+
},
|
| 113 |
+
{
|
| 114 |
+
"epoch": 0.35,
|
| 115 |
+
"learning_rate": 4.826923076923077e-05,
|
| 116 |
+
"loss": 3.756,
|
| 117 |
+
"step": 108
|
| 118 |
+
},
|
| 119 |
+
{
|
| 120 |
+
"epoch": 0.37,
|
| 121 |
+
"learning_rate": 4.8173076923076925e-05,
|
| 122 |
+
"loss": 3.7439,
|
| 123 |
+
"step": 114
|
| 124 |
+
},
|
| 125 |
+
{
|
| 126 |
+
"epoch": 0.38,
|
| 127 |
+
"learning_rate": 4.8076923076923084e-05,
|
| 128 |
+
"loss": 3.761,
|
| 129 |
+
"step": 120
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.4,
|
| 133 |
+
"learning_rate": 4.7980769230769236e-05,
|
| 134 |
+
"loss": 3.7318,
|
| 135 |
+
"step": 126
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.42,
|
| 139 |
+
"learning_rate": 4.788461538461539e-05,
|
| 140 |
+
"loss": 3.745,
|
| 141 |
+
"step": 132
|
| 142 |
+
},
|
| 143 |
+
{
|
| 144 |
+
"epoch": 0.44,
|
| 145 |
+
"learning_rate": 4.778846153846154e-05,
|
| 146 |
+
"loss": 3.784,
|
| 147 |
+
"step": 138
|
| 148 |
+
},
|
| 149 |
+
{
|
| 150 |
+
"epoch": 0.46,
|
| 151 |
+
"learning_rate": 4.76923076923077e-05,
|
| 152 |
+
"loss": 3.6114,
|
| 153 |
+
"step": 144
|
| 154 |
+
},
|
| 155 |
+
{
|
| 156 |
+
"epoch": 0.48,
|
| 157 |
+
"learning_rate": 4.7596153846153844e-05,
|
| 158 |
+
"loss": 3.637,
|
| 159 |
+
"step": 150
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 0.5,
|
| 163 |
+
"learning_rate": 4.75e-05,
|
| 164 |
+
"loss": 3.5881,
|
| 165 |
+
"step": 156
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 0.52,
|
| 169 |
+
"learning_rate": 4.7403846153846155e-05,
|
| 170 |
+
"loss": 3.5336,
|
| 171 |
+
"step": 162
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 0.54,
|
| 175 |
+
"learning_rate": 4.730769230769231e-05,
|
| 176 |
+
"loss": 3.5313,
|
| 177 |
+
"step": 168
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.56,
|
| 181 |
+
"learning_rate": 4.7211538461538465e-05,
|
| 182 |
+
"loss": 3.6357,
|
| 183 |
+
"step": 174
|
| 184 |
+
},
|
| 185 |
+
{
|
| 186 |
+
"epoch": 0.58,
|
| 187 |
+
"learning_rate": 4.711538461538462e-05,
|
| 188 |
+
"loss": 3.5818,
|
| 189 |
+
"step": 180
|
| 190 |
+
},
|
| 191 |
+
{
|
| 192 |
+
"epoch": 0.6,
|
| 193 |
+
"learning_rate": 4.701923076923077e-05,
|
| 194 |
+
"loss": 3.5017,
|
| 195 |
+
"step": 186
|
| 196 |
+
},
|
| 197 |
+
{
|
| 198 |
+
"epoch": 0.62,
|
| 199 |
+
"learning_rate": 4.692307692307693e-05,
|
| 200 |
+
"loss": 3.5815,
|
| 201 |
+
"step": 192
|
| 202 |
+
},
|
| 203 |
+
{
|
| 204 |
+
"epoch": 0.63,
|
| 205 |
+
"learning_rate": 4.682692307692308e-05,
|
| 206 |
+
"loss": 3.611,
|
| 207 |
+
"step": 198
|
| 208 |
+
},
|
| 209 |
+
{
|
| 210 |
+
"epoch": 0.65,
|
| 211 |
+
"learning_rate": 4.673076923076923e-05,
|
| 212 |
+
"loss": 3.6096,
|
| 213 |
+
"step": 204
|
| 214 |
+
},
|
| 215 |
+
{
|
| 216 |
+
"epoch": 0.67,
|
| 217 |
+
"learning_rate": 4.6634615384615384e-05,
|
| 218 |
+
"loss": 3.5436,
|
| 219 |
+
"step": 210
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.69,
|
| 223 |
+
"learning_rate": 4.653846153846154e-05,
|
| 224 |
+
"loss": 3.5258,
|
| 225 |
+
"step": 216
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 0.71,
|
| 229 |
+
"learning_rate": 4.6442307692307695e-05,
|
| 230 |
+
"loss": 3.5277,
|
| 231 |
+
"step": 222
|
| 232 |
+
},
|
| 233 |
+
{
|
| 234 |
+
"epoch": 0.73,
|
| 235 |
+
"learning_rate": 4.634615384615385e-05,
|
| 236 |
+
"loss": 3.505,
|
| 237 |
+
"step": 228
|
| 238 |
+
},
|
| 239 |
+
{
|
| 240 |
+
"epoch": 0.75,
|
| 241 |
+
"learning_rate": 4.6250000000000006e-05,
|
| 242 |
+
"loss": 3.4665,
|
| 243 |
+
"step": 234
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"epoch": 0.77,
|
| 247 |
+
"learning_rate": 4.615384615384616e-05,
|
| 248 |
+
"loss": 3.4245,
|
| 249 |
+
"step": 240
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 0.79,
|
| 253 |
+
"learning_rate": 4.605769230769231e-05,
|
| 254 |
+
"loss": 3.4547,
|
| 255 |
+
"step": 246
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 0.81,
|
| 259 |
+
"learning_rate": 4.596153846153846e-05,
|
| 260 |
+
"loss": 3.4382,
|
| 261 |
+
"step": 252
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.83,
|
| 265 |
+
"learning_rate": 4.5865384615384614e-05,
|
| 266 |
+
"loss": 3.482,
|
| 267 |
+
"step": 258
|
| 268 |
+
},
|
| 269 |
+
{
|
| 270 |
+
"epoch": 0.85,
|
| 271 |
+
"learning_rate": 4.576923076923077e-05,
|
| 272 |
+
"loss": 3.4795,
|
| 273 |
+
"step": 264
|
| 274 |
+
},
|
| 275 |
+
{
|
| 276 |
+
"epoch": 0.87,
|
| 277 |
+
"learning_rate": 4.5673076923076925e-05,
|
| 278 |
+
"loss": 3.5018,
|
| 279 |
+
"step": 270
|
| 280 |
+
},
|
| 281 |
+
{
|
| 282 |
+
"epoch": 0.88,
|
| 283 |
+
"learning_rate": 4.557692307692308e-05,
|
| 284 |
+
"loss": 3.448,
|
| 285 |
+
"step": 276
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 0.9,
|
| 289 |
+
"learning_rate": 4.548076923076923e-05,
|
| 290 |
+
"loss": 3.4206,
|
| 291 |
+
"step": 282
|
| 292 |
+
},
|
| 293 |
+
{
|
| 294 |
+
"epoch": 0.92,
|
| 295 |
+
"learning_rate": 4.538461538461539e-05,
|
| 296 |
+
"loss": 3.528,
|
| 297 |
+
"step": 288
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 0.94,
|
| 301 |
+
"learning_rate": 4.528846153846154e-05,
|
| 302 |
+
"loss": 3.475,
|
| 303 |
+
"step": 294
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.96,
|
| 307 |
+
"learning_rate": 4.519230769230769e-05,
|
| 308 |
+
"loss": 3.4142,
|
| 309 |
+
"step": 300
|
| 310 |
+
},
|
| 311 |
+
{
|
| 312 |
+
"epoch": 0.98,
|
| 313 |
+
"learning_rate": 4.509615384615385e-05,
|
| 314 |
+
"loss": 3.4156,
|
| 315 |
+
"step": 306
|
| 316 |
+
},
|
| 317 |
+
{
|
| 318 |
+
"epoch": 1.0,
|
| 319 |
+
"learning_rate": 4.5e-05,
|
| 320 |
+
"loss": 3.3274,
|
| 321 |
+
"step": 312
|
| 322 |
+
},
|
| 323 |
+
{
|
| 324 |
+
"epoch": 1.02,
|
| 325 |
+
"learning_rate": 4.4903846153846155e-05,
|
| 326 |
+
"loss": 3.3594,
|
| 327 |
+
"step": 318
|
| 328 |
+
},
|
| 329 |
+
{
|
| 330 |
+
"epoch": 1.04,
|
| 331 |
+
"learning_rate": 4.4807692307692314e-05,
|
| 332 |
+
"loss": 3.2517,
|
| 333 |
+
"step": 324
|
| 334 |
+
},
|
| 335 |
+
{
|
| 336 |
+
"epoch": 1.06,
|
| 337 |
+
"learning_rate": 4.4711538461538466e-05,
|
| 338 |
+
"loss": 3.2934,
|
| 339 |
+
"step": 330
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 1.08,
|
| 343 |
+
"learning_rate": 4.461538461538462e-05,
|
| 344 |
+
"loss": 3.2986,
|
| 345 |
+
"step": 336
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 1.1,
|
| 349 |
+
"learning_rate": 4.451923076923077e-05,
|
| 350 |
+
"loss": 3.2655,
|
| 351 |
+
"step": 342
|
| 352 |
+
},
|
| 353 |
+
{
|
| 354 |
+
"epoch": 1.12,
|
| 355 |
+
"learning_rate": 4.442307692307692e-05,
|
| 356 |
+
"loss": 3.4058,
|
| 357 |
+
"step": 348
|
| 358 |
+
},
|
| 359 |
+
{
|
| 360 |
+
"epoch": 1.13,
|
| 361 |
+
"learning_rate": 4.4326923076923074e-05,
|
| 362 |
+
"loss": 3.265,
|
| 363 |
+
"step": 354
|
| 364 |
+
},
|
| 365 |
+
{
|
| 366 |
+
"epoch": 1.15,
|
| 367 |
+
"learning_rate": 4.423076923076923e-05,
|
| 368 |
+
"loss": 3.3208,
|
| 369 |
+
"step": 360
|
| 370 |
+
},
|
| 371 |
+
{
|
| 372 |
+
"epoch": 1.17,
|
| 373 |
+
"learning_rate": 4.4134615384615385e-05,
|
| 374 |
+
"loss": 3.238,
|
| 375 |
+
"step": 366
|
| 376 |
+
},
|
| 377 |
+
{
|
| 378 |
+
"epoch": 1.19,
|
| 379 |
+
"learning_rate": 4.403846153846154e-05,
|
| 380 |
+
"loss": 3.2153,
|
| 381 |
+
"step": 372
|
| 382 |
+
},
|
| 383 |
+
{
|
| 384 |
+
"epoch": 1.21,
|
| 385 |
+
"learning_rate": 4.3942307692307695e-05,
|
| 386 |
+
"loss": 3.264,
|
| 387 |
+
"step": 378
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 1.23,
|
| 391 |
+
"learning_rate": 4.384615384615385e-05,
|
| 392 |
+
"loss": 3.2911,
|
| 393 |
+
"step": 384
|
| 394 |
+
},
|
| 395 |
+
{
|
| 396 |
+
"epoch": 1.25,
|
| 397 |
+
"learning_rate": 4.375e-05,
|
| 398 |
+
"loss": 3.3027,
|
| 399 |
+
"step": 390
|
| 400 |
+
},
|
| 401 |
+
{
|
| 402 |
+
"epoch": 1.27,
|
| 403 |
+
"learning_rate": 4.365384615384616e-05,
|
| 404 |
+
"loss": 3.2589,
|
| 405 |
+
"step": 396
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 1.29,
|
| 409 |
+
"learning_rate": 4.355769230769231e-05,
|
| 410 |
+
"loss": 3.3683,
|
| 411 |
+
"step": 402
|
| 412 |
+
},
|
| 413 |
+
{
|
| 414 |
+
"epoch": 1.31,
|
| 415 |
+
"learning_rate": 4.346153846153846e-05,
|
| 416 |
+
"loss": 3.2849,
|
| 417 |
+
"step": 408
|
| 418 |
+
},
|
| 419 |
+
{
|
| 420 |
+
"epoch": 1.33,
|
| 421 |
+
"learning_rate": 4.336538461538462e-05,
|
| 422 |
+
"loss": 3.2397,
|
| 423 |
+
"step": 414
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 1.35,
|
| 427 |
+
"learning_rate": 4.326923076923077e-05,
|
| 428 |
+
"loss": 3.2128,
|
| 429 |
+
"step": 420
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 1.37,
|
| 433 |
+
"learning_rate": 4.3173076923076925e-05,
|
| 434 |
+
"loss": 3.1944,
|
| 435 |
+
"step": 426
|
| 436 |
+
},
|
| 437 |
+
{
|
| 438 |
+
"epoch": 1.38,
|
| 439 |
+
"learning_rate": 4.3076923076923084e-05,
|
| 440 |
+
"loss": 3.1837,
|
| 441 |
+
"step": 432
|
| 442 |
+
},
|
| 443 |
+
{
|
| 444 |
+
"epoch": 1.4,
|
| 445 |
+
"learning_rate": 4.2980769230769236e-05,
|
| 446 |
+
"loss": 3.1793,
|
| 447 |
+
"step": 438
|
| 448 |
+
},
|
| 449 |
+
{
|
| 450 |
+
"epoch": 1.42,
|
| 451 |
+
"learning_rate": 4.288461538461538e-05,
|
| 452 |
+
"loss": 3.1447,
|
| 453 |
+
"step": 444
|
| 454 |
+
},
|
| 455 |
+
{
|
| 456 |
+
"epoch": 1.44,
|
| 457 |
+
"learning_rate": 4.278846153846154e-05,
|
| 458 |
+
"loss": 3.1028,
|
| 459 |
+
"step": 450
|
| 460 |
+
},
|
| 461 |
+
{
|
| 462 |
+
"epoch": 1.46,
|
| 463 |
+
"learning_rate": 4.269230769230769e-05,
|
| 464 |
+
"loss": 3.2471,
|
| 465 |
+
"step": 456
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 1.48,
|
| 469 |
+
"learning_rate": 4.2596153846153844e-05,
|
| 470 |
+
"loss": 3.1855,
|
| 471 |
+
"step": 462
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 1.5,
|
| 475 |
+
"learning_rate": 4.25e-05,
|
| 476 |
+
"loss": 3.1817,
|
| 477 |
+
"step": 468
|
| 478 |
+
},
|
| 479 |
+
{
|
| 480 |
+
"epoch": 1.52,
|
| 481 |
+
"learning_rate": 4.2403846153846155e-05,
|
| 482 |
+
"loss": 3.214,
|
| 483 |
+
"step": 474
|
| 484 |
+
},
|
| 485 |
+
{
|
| 486 |
+
"epoch": 1.54,
|
| 487 |
+
"learning_rate": 4.230769230769231e-05,
|
| 488 |
+
"loss": 2.997,
|
| 489 |
+
"step": 480
|
| 490 |
+
},
|
| 491 |
+
{
|
| 492 |
+
"epoch": 1.56,
|
| 493 |
+
"learning_rate": 4.2211538461538466e-05,
|
| 494 |
+
"loss": 3.2059,
|
| 495 |
+
"step": 486
|
| 496 |
+
},
|
| 497 |
+
{
|
| 498 |
+
"epoch": 1.58,
|
| 499 |
+
"learning_rate": 4.211538461538462e-05,
|
| 500 |
+
"loss": 3.1517,
|
| 501 |
+
"step": 492
|
| 502 |
+
},
|
| 503 |
+
{
|
| 504 |
+
"epoch": 1.6,
|
| 505 |
+
"learning_rate": 4.201923076923077e-05,
|
| 506 |
+
"loss": 3.2236,
|
| 507 |
+
"step": 498
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 1.62,
|
| 511 |
+
"learning_rate": 4.192307692307693e-05,
|
| 512 |
+
"loss": 3.1939,
|
| 513 |
+
"step": 504
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 1.63,
|
| 517 |
+
"learning_rate": 4.182692307692308e-05,
|
| 518 |
+
"loss": 3.1143,
|
| 519 |
+
"step": 510
|
| 520 |
+
},
|
| 521 |
+
{
|
| 522 |
+
"epoch": 1.65,
|
| 523 |
+
"learning_rate": 4.173076923076923e-05,
|
| 524 |
+
"loss": 3.2261,
|
| 525 |
+
"step": 516
|
| 526 |
+
},
|
| 527 |
+
{
|
| 528 |
+
"epoch": 1.67,
|
| 529 |
+
"learning_rate": 4.163461538461539e-05,
|
| 530 |
+
"loss": 3.2095,
|
| 531 |
+
"step": 522
|
| 532 |
+
},
|
| 533 |
+
{
|
| 534 |
+
"epoch": 1.69,
|
| 535 |
+
"learning_rate": 4.1538461538461544e-05,
|
| 536 |
+
"loss": 3.2498,
|
| 537 |
+
"step": 528
|
| 538 |
+
},
|
| 539 |
+
{
|
| 540 |
+
"epoch": 1.71,
|
| 541 |
+
"learning_rate": 4.1442307692307696e-05,
|
| 542 |
+
"loss": 3.1493,
|
| 543 |
+
"step": 534
|
| 544 |
+
},
|
| 545 |
+
{
|
| 546 |
+
"epoch": 1.73,
|
| 547 |
+
"learning_rate": 4.134615384615385e-05,
|
| 548 |
+
"loss": 3.0919,
|
| 549 |
+
"step": 540
|
| 550 |
+
},
|
| 551 |
+
{
|
| 552 |
+
"epoch": 1.75,
|
| 553 |
+
"learning_rate": 4.125e-05,
|
| 554 |
+
"loss": 3.2617,
|
| 555 |
+
"step": 546
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 1.77,
|
| 559 |
+
"learning_rate": 4.115384615384615e-05,
|
| 560 |
+
"loss": 3.1447,
|
| 561 |
+
"step": 552
|
| 562 |
+
},
|
| 563 |
+
{
|
| 564 |
+
"epoch": 1.79,
|
| 565 |
+
"learning_rate": 4.105769230769231e-05,
|
| 566 |
+
"loss": 3.2025,
|
| 567 |
+
"step": 558
|
| 568 |
+
},
|
| 569 |
+
{
|
| 570 |
+
"epoch": 1.81,
|
| 571 |
+
"learning_rate": 4.096153846153846e-05,
|
| 572 |
+
"loss": 3.1829,
|
| 573 |
+
"step": 564
|
| 574 |
+
},
|
| 575 |
+
{
|
| 576 |
+
"epoch": 1.83,
|
| 577 |
+
"learning_rate": 4.0865384615384615e-05,
|
| 578 |
+
"loss": 3.2216,
|
| 579 |
+
"step": 570
|
| 580 |
+
},
|
| 581 |
+
{
|
| 582 |
+
"epoch": 1.85,
|
| 583 |
+
"learning_rate": 4.0769230769230773e-05,
|
| 584 |
+
"loss": 3.1391,
|
| 585 |
+
"step": 576
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 1.87,
|
| 589 |
+
"learning_rate": 4.0673076923076926e-05,
|
| 590 |
+
"loss": 3.1344,
|
| 591 |
+
"step": 582
|
| 592 |
+
},
|
| 593 |
+
{
|
| 594 |
+
"epoch": 1.88,
|
| 595 |
+
"learning_rate": 4.057692307692308e-05,
|
| 596 |
+
"loss": 3.1699,
|
| 597 |
+
"step": 588
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 1.9,
|
| 601 |
+
"learning_rate": 4.0480769230769236e-05,
|
| 602 |
+
"loss": 3.0138,
|
| 603 |
+
"step": 594
|
| 604 |
+
},
|
| 605 |
+
{
|
| 606 |
+
"epoch": 1.92,
|
| 607 |
+
"learning_rate": 4.038461538461539e-05,
|
| 608 |
+
"loss": 3.1157,
|
| 609 |
+
"step": 600
|
| 610 |
+
},
|
| 611 |
+
{
|
| 612 |
+
"epoch": 1.94,
|
| 613 |
+
"learning_rate": 4.028846153846154e-05,
|
| 614 |
+
"loss": 3.1173,
|
| 615 |
+
"step": 606
|
| 616 |
+
},
|
| 617 |
+
{
|
| 618 |
+
"epoch": 1.96,
|
| 619 |
+
"learning_rate": 4.019230769230769e-05,
|
| 620 |
+
"loss": 3.2019,
|
| 621 |
+
"step": 612
|
| 622 |
+
},
|
| 623 |
+
{
|
| 624 |
+
"epoch": 1.98,
|
| 625 |
+
"learning_rate": 4.009615384615385e-05,
|
| 626 |
+
"loss": 3.0801,
|
| 627 |
+
"step": 618
|
| 628 |
+
},
|
| 629 |
+
{
|
| 630 |
+
"epoch": 2.0,
|
| 631 |
+
"learning_rate": 4e-05,
|
| 632 |
+
"loss": 3.0888,
|
| 633 |
+
"step": 624
|
| 634 |
+
},
|
| 635 |
+
{
|
| 636 |
+
"epoch": 2.02,
|
| 637 |
+
"learning_rate": 3.9903846153846155e-05,
|
| 638 |
+
"loss": 3.0332,
|
| 639 |
+
"step": 630
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 2.04,
|
| 643 |
+
"learning_rate": 3.980769230769231e-05,
|
| 644 |
+
"loss": 2.9418,
|
| 645 |
+
"step": 636
|
| 646 |
+
},
|
| 647 |
+
{
|
| 648 |
+
"epoch": 2.06,
|
| 649 |
+
"learning_rate": 3.971153846153846e-05,
|
| 650 |
+
"loss": 2.933,
|
| 651 |
+
"step": 642
|
| 652 |
+
},
|
| 653 |
+
{
|
| 654 |
+
"epoch": 2.08,
|
| 655 |
+
"learning_rate": 3.961538461538462e-05,
|
| 656 |
+
"loss": 3.0467,
|
| 657 |
+
"step": 648
|
| 658 |
+
},
|
| 659 |
+
{
|
| 660 |
+
"epoch": 2.1,
|
| 661 |
+
"learning_rate": 3.951923076923077e-05,
|
| 662 |
+
"loss": 3.0991,
|
| 663 |
+
"step": 654
|
| 664 |
+
},
|
| 665 |
+
{
|
| 666 |
+
"epoch": 2.12,
|
| 667 |
+
"learning_rate": 3.942307692307692e-05,
|
| 668 |
+
"loss": 2.913,
|
| 669 |
+
"step": 660
|
| 670 |
+
},
|
| 671 |
+
{
|
| 672 |
+
"epoch": 2.13,
|
| 673 |
+
"learning_rate": 3.932692307692308e-05,
|
| 674 |
+
"loss": 3.0531,
|
| 675 |
+
"step": 666
|
| 676 |
+
},
|
| 677 |
+
{
|
| 678 |
+
"epoch": 2.15,
|
| 679 |
+
"learning_rate": 3.923076923076923e-05,
|
| 680 |
+
"loss": 2.9838,
|
| 681 |
+
"step": 672
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 2.17,
|
| 685 |
+
"learning_rate": 3.9134615384615385e-05,
|
| 686 |
+
"loss": 2.9406,
|
| 687 |
+
"step": 678
|
| 688 |
+
},
|
| 689 |
+
{
|
| 690 |
+
"epoch": 2.19,
|
| 691 |
+
"learning_rate": 3.903846153846154e-05,
|
| 692 |
+
"loss": 3.046,
|
| 693 |
+
"step": 684
|
| 694 |
+
},
|
| 695 |
+
{
|
| 696 |
+
"epoch": 2.21,
|
| 697 |
+
"learning_rate": 3.8942307692307696e-05,
|
| 698 |
+
"loss": 3.006,
|
| 699 |
+
"step": 690
|
| 700 |
+
},
|
| 701 |
+
{
|
| 702 |
+
"epoch": 2.23,
|
| 703 |
+
"learning_rate": 3.884615384615385e-05,
|
| 704 |
+
"loss": 2.9774,
|
| 705 |
+
"step": 696
|
| 706 |
+
},
|
| 707 |
+
{
|
| 708 |
+
"epoch": 2.25,
|
| 709 |
+
"learning_rate": 3.875e-05,
|
| 710 |
+
"loss": 2.9937,
|
| 711 |
+
"step": 702
|
| 712 |
+
},
|
| 713 |
+
{
|
| 714 |
+
"epoch": 2.27,
|
| 715 |
+
"learning_rate": 3.865384615384616e-05,
|
| 716 |
+
"loss": 2.9737,
|
| 717 |
+
"step": 708
|
| 718 |
+
},
|
| 719 |
+
{
|
| 720 |
+
"epoch": 2.29,
|
| 721 |
+
"learning_rate": 3.855769230769231e-05,
|
| 722 |
+
"loss": 2.9772,
|
| 723 |
+
"step": 714
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 2.31,
|
| 727 |
+
"learning_rate": 3.846153846153846e-05,
|
| 728 |
+
"loss": 2.8405,
|
| 729 |
+
"step": 720
|
| 730 |
+
},
|
| 731 |
+
{
|
| 732 |
+
"epoch": 2.33,
|
| 733 |
+
"learning_rate": 3.836538461538462e-05,
|
| 734 |
+
"loss": 3.0856,
|
| 735 |
+
"step": 726
|
| 736 |
+
},
|
| 737 |
+
{
|
| 738 |
+
"epoch": 2.35,
|
| 739 |
+
"learning_rate": 3.826923076923077e-05,
|
| 740 |
+
"loss": 2.9947,
|
| 741 |
+
"step": 732
|
| 742 |
+
},
|
| 743 |
+
{
|
| 744 |
+
"epoch": 2.37,
|
| 745 |
+
"learning_rate": 3.8173076923076926e-05,
|
| 746 |
+
"loss": 2.9251,
|
| 747 |
+
"step": 738
|
| 748 |
+
},
|
| 749 |
+
{
|
| 750 |
+
"epoch": 2.38,
|
| 751 |
+
"learning_rate": 3.807692307692308e-05,
|
| 752 |
+
"loss": 2.9613,
|
| 753 |
+
"step": 744
|
| 754 |
+
},
|
| 755 |
+
{
|
| 756 |
+
"epoch": 2.4,
|
| 757 |
+
"learning_rate": 3.798076923076923e-05,
|
| 758 |
+
"loss": 2.9402,
|
| 759 |
+
"step": 750
|
| 760 |
+
},
|
| 761 |
+
{
|
| 762 |
+
"epoch": 2.42,
|
| 763 |
+
"learning_rate": 3.788461538461538e-05,
|
| 764 |
+
"loss": 2.8823,
|
| 765 |
+
"step": 756
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 2.44,
|
| 769 |
+
"learning_rate": 3.778846153846154e-05,
|
| 770 |
+
"loss": 2.8906,
|
| 771 |
+
"step": 762
|
| 772 |
+
},
|
| 773 |
+
{
|
| 774 |
+
"epoch": 2.46,
|
| 775 |
+
"learning_rate": 3.769230769230769e-05,
|
| 776 |
+
"loss": 2.9475,
|
| 777 |
+
"step": 768
|
| 778 |
+
},
|
| 779 |
+
{
|
| 780 |
+
"epoch": 2.48,
|
| 781 |
+
"learning_rate": 3.7596153846153845e-05,
|
| 782 |
+
"loss": 2.9999,
|
| 783 |
+
"step": 774
|
| 784 |
+
},
|
| 785 |
+
{
|
| 786 |
+
"epoch": 2.5,
|
| 787 |
+
"learning_rate": 3.7500000000000003e-05,
|
| 788 |
+
"loss": 2.9503,
|
| 789 |
+
"step": 780
|
| 790 |
+
},
|
| 791 |
+
{
|
| 792 |
+
"epoch": 2.52,
|
| 793 |
+
"learning_rate": 3.7403846153846156e-05,
|
| 794 |
+
"loss": 3.0112,
|
| 795 |
+
"step": 786
|
| 796 |
+
},
|
| 797 |
+
{
|
| 798 |
+
"epoch": 2.54,
|
| 799 |
+
"learning_rate": 3.730769230769231e-05,
|
| 800 |
+
"loss": 3.0096,
|
| 801 |
+
"step": 792
|
| 802 |
+
},
|
| 803 |
+
{
|
| 804 |
+
"epoch": 2.56,
|
| 805 |
+
"learning_rate": 3.7211538461538466e-05,
|
| 806 |
+
"loss": 2.9804,
|
| 807 |
+
"step": 798
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 2.58,
|
| 811 |
+
"learning_rate": 3.711538461538462e-05,
|
| 812 |
+
"loss": 2.9203,
|
| 813 |
+
"step": 804
|
| 814 |
+
},
|
| 815 |
+
{
|
| 816 |
+
"epoch": 2.6,
|
| 817 |
+
"learning_rate": 3.701923076923077e-05,
|
| 818 |
+
"loss": 2.939,
|
| 819 |
+
"step": 810
|
| 820 |
+
},
|
| 821 |
+
{
|
| 822 |
+
"epoch": 2.62,
|
| 823 |
+
"learning_rate": 3.692307692307693e-05,
|
| 824 |
+
"loss": 2.8455,
|
| 825 |
+
"step": 816
|
| 826 |
+
},
|
| 827 |
+
{
|
| 828 |
+
"epoch": 2.63,
|
| 829 |
+
"learning_rate": 3.682692307692308e-05,
|
| 830 |
+
"loss": 2.9651,
|
| 831 |
+
"step": 822
|
| 832 |
+
},
|
| 833 |
+
{
|
| 834 |
+
"epoch": 2.65,
|
| 835 |
+
"learning_rate": 3.673076923076923e-05,
|
| 836 |
+
"loss": 2.9528,
|
| 837 |
+
"step": 828
|
| 838 |
+
},
|
| 839 |
+
{
|
| 840 |
+
"epoch": 2.67,
|
| 841 |
+
"learning_rate": 3.6634615384615385e-05,
|
| 842 |
+
"loss": 2.8042,
|
| 843 |
+
"step": 834
|
| 844 |
+
},
|
| 845 |
+
{
|
| 846 |
+
"epoch": 2.69,
|
| 847 |
+
"learning_rate": 3.653846153846154e-05,
|
| 848 |
+
"loss": 2.8311,
|
| 849 |
+
"step": 840
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 2.71,
|
| 853 |
+
"learning_rate": 3.644230769230769e-05,
|
| 854 |
+
"loss": 2.8888,
|
| 855 |
+
"step": 846
|
| 856 |
+
},
|
| 857 |
+
{
|
| 858 |
+
"epoch": 2.73,
|
| 859 |
+
"learning_rate": 3.634615384615385e-05,
|
| 860 |
+
"loss": 2.9151,
|
| 861 |
+
"step": 852
|
| 862 |
+
},
|
| 863 |
+
{
|
| 864 |
+
"epoch": 2.75,
|
| 865 |
+
"learning_rate": 3.625e-05,
|
| 866 |
+
"loss": 2.9463,
|
| 867 |
+
"step": 858
|
| 868 |
+
},
|
| 869 |
+
{
|
| 870 |
+
"epoch": 2.77,
|
| 871 |
+
"learning_rate": 3.615384615384615e-05,
|
| 872 |
+
"loss": 2.957,
|
| 873 |
+
"step": 864
|
| 874 |
+
},
|
| 875 |
+
{
|
| 876 |
+
"epoch": 2.79,
|
| 877 |
+
"learning_rate": 3.605769230769231e-05,
|
| 878 |
+
"loss": 2.9473,
|
| 879 |
+
"step": 870
|
| 880 |
+
},
|
| 881 |
+
{
|
| 882 |
+
"epoch": 2.81,
|
| 883 |
+
"learning_rate": 3.596153846153846e-05,
|
| 884 |
+
"loss": 2.9994,
|
| 885 |
+
"step": 876
|
| 886 |
+
},
|
| 887 |
+
{
|
| 888 |
+
"epoch": 2.83,
|
| 889 |
+
"learning_rate": 3.5865384615384615e-05,
|
| 890 |
+
"loss": 3.0486,
|
| 891 |
+
"step": 882
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 2.85,
|
| 895 |
+
"learning_rate": 3.5769230769230774e-05,
|
| 896 |
+
"loss": 2.9487,
|
| 897 |
+
"step": 888
|
| 898 |
+
},
|
| 899 |
+
{
|
| 900 |
+
"epoch": 2.87,
|
| 901 |
+
"learning_rate": 3.5673076923076926e-05,
|
| 902 |
+
"loss": 3.0173,
|
| 903 |
+
"step": 894
|
| 904 |
+
},
|
| 905 |
+
{
|
| 906 |
+
"epoch": 2.88,
|
| 907 |
+
"learning_rate": 3.557692307692308e-05,
|
| 908 |
+
"loss": 2.8656,
|
| 909 |
+
"step": 900
|
| 910 |
+
},
|
| 911 |
+
{
|
| 912 |
+
"epoch": 2.9,
|
| 913 |
+
"learning_rate": 3.548076923076924e-05,
|
| 914 |
+
"loss": 2.8834,
|
| 915 |
+
"step": 906
|
| 916 |
+
},
|
| 917 |
+
{
|
| 918 |
+
"epoch": 2.92,
|
| 919 |
+
"learning_rate": 3.538461538461539e-05,
|
| 920 |
+
"loss": 2.9829,
|
| 921 |
+
"step": 912
|
| 922 |
+
},
|
| 923 |
+
{
|
| 924 |
+
"epoch": 2.94,
|
| 925 |
+
"learning_rate": 3.528846153846154e-05,
|
| 926 |
+
"loss": 3.001,
|
| 927 |
+
"step": 918
|
| 928 |
+
},
|
| 929 |
+
{
|
| 930 |
+
"epoch": 2.96,
|
| 931 |
+
"learning_rate": 3.51923076923077e-05,
|
| 932 |
+
"loss": 2.9618,
|
| 933 |
+
"step": 924
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 2.98,
|
| 937 |
+
"learning_rate": 3.5096153846153845e-05,
|
| 938 |
+
"loss": 2.7964,
|
| 939 |
+
"step": 930
|
| 940 |
+
},
|
| 941 |
+
{
|
| 942 |
+
"epoch": 3.0,
|
| 943 |
+
"learning_rate": 3.5e-05,
|
| 944 |
+
"loss": 2.89,
|
| 945 |
+
"step": 936
|
| 946 |
+
},
|
| 947 |
+
{
|
| 948 |
+
"epoch": 3.02,
|
| 949 |
+
"learning_rate": 3.4903846153846156e-05,
|
| 950 |
+
"loss": 2.7191,
|
| 951 |
+
"step": 942
|
| 952 |
+
},
|
| 953 |
+
{
|
| 954 |
+
"epoch": 3.04,
|
| 955 |
+
"learning_rate": 3.480769230769231e-05,
|
| 956 |
+
"loss": 2.7875,
|
| 957 |
+
"step": 948
|
| 958 |
+
},
|
| 959 |
+
{
|
| 960 |
+
"epoch": 3.06,
|
| 961 |
+
"learning_rate": 3.471153846153846e-05,
|
| 962 |
+
"loss": 2.8498,
|
| 963 |
+
"step": 954
|
| 964 |
+
},
|
| 965 |
+
{
|
| 966 |
+
"epoch": 3.08,
|
| 967 |
+
"learning_rate": 3.461538461538462e-05,
|
| 968 |
+
"loss": 2.7858,
|
| 969 |
+
"step": 960
|
| 970 |
+
},
|
| 971 |
+
{
|
| 972 |
+
"epoch": 3.1,
|
| 973 |
+
"learning_rate": 3.451923076923077e-05,
|
| 974 |
+
"loss": 2.7435,
|
| 975 |
+
"step": 966
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 3.12,
|
| 979 |
+
"learning_rate": 3.442307692307692e-05,
|
| 980 |
+
"loss": 2.8449,
|
| 981 |
+
"step": 972
|
| 982 |
+
},
|
| 983 |
+
{
|
| 984 |
+
"epoch": 3.13,
|
| 985 |
+
"learning_rate": 3.432692307692308e-05,
|
| 986 |
+
"loss": 2.7817,
|
| 987 |
+
"step": 978
|
| 988 |
+
},
|
| 989 |
+
{
|
| 990 |
+
"epoch": 3.15,
|
| 991 |
+
"learning_rate": 3.4230769230769234e-05,
|
| 992 |
+
"loss": 2.8797,
|
| 993 |
+
"step": 984
|
| 994 |
+
},
|
| 995 |
+
{
|
| 996 |
+
"epoch": 3.17,
|
| 997 |
+
"learning_rate": 3.4134615384615386e-05,
|
| 998 |
+
"loss": 2.7719,
|
| 999 |
+
"step": 990
|
| 1000 |
+
},
|
| 1001 |
+
{
|
| 1002 |
+
"epoch": 3.19,
|
| 1003 |
+
"learning_rate": 3.4038461538461544e-05,
|
| 1004 |
+
"loss": 2.7919,
|
| 1005 |
+
"step": 996
|
| 1006 |
+
},
|
| 1007 |
+
{
|
| 1008 |
+
"epoch": 3.21,
|
| 1009 |
+
"learning_rate": 3.3942307692307696e-05,
|
| 1010 |
+
"loss": 2.8072,
|
| 1011 |
+
"step": 1002
|
| 1012 |
+
},
|
| 1013 |
+
{
|
| 1014 |
+
"epoch": 3.23,
|
| 1015 |
+
"learning_rate": 3.384615384615385e-05,
|
| 1016 |
+
"loss": 2.7281,
|
| 1017 |
+
"step": 1008
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 3.25,
|
| 1021 |
+
"learning_rate": 3.375000000000001e-05,
|
| 1022 |
+
"loss": 2.8431,
|
| 1023 |
+
"step": 1014
|
| 1024 |
+
},
|
| 1025 |
+
{
|
| 1026 |
+
"epoch": 3.27,
|
| 1027 |
+
"learning_rate": 3.365384615384616e-05,
|
| 1028 |
+
"loss": 2.7154,
|
| 1029 |
+
"step": 1020
|
| 1030 |
+
},
|
| 1031 |
+
{
|
| 1032 |
+
"epoch": 3.29,
|
| 1033 |
+
"learning_rate": 3.3557692307692304e-05,
|
| 1034 |
+
"loss": 2.83,
|
| 1035 |
+
"step": 1026
|
| 1036 |
+
},
|
| 1037 |
+
{
|
| 1038 |
+
"epoch": 3.31,
|
| 1039 |
+
"learning_rate": 3.346153846153846e-05,
|
| 1040 |
+
"loss": 2.8434,
|
| 1041 |
+
"step": 1032
|
| 1042 |
+
},
|
| 1043 |
+
{
|
| 1044 |
+
"epoch": 3.33,
|
| 1045 |
+
"learning_rate": 3.3365384615384615e-05,
|
| 1046 |
+
"loss": 2.8399,
|
| 1047 |
+
"step": 1038
|
| 1048 |
+
},
|
| 1049 |
+
{
|
| 1050 |
+
"epoch": 3.35,
|
| 1051 |
+
"learning_rate": 3.326923076923077e-05,
|
| 1052 |
+
"loss": 2.7072,
|
| 1053 |
+
"step": 1044
|
| 1054 |
+
},
|
| 1055 |
+
{
|
| 1056 |
+
"epoch": 3.37,
|
| 1057 |
+
"learning_rate": 3.3173076923076926e-05,
|
| 1058 |
+
"loss": 2.773,
|
| 1059 |
+
"step": 1050
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 3.38,
|
| 1063 |
+
"learning_rate": 3.307692307692308e-05,
|
| 1064 |
+
"loss": 2.822,
|
| 1065 |
+
"step": 1056
|
| 1066 |
+
},
|
| 1067 |
+
{
|
| 1068 |
+
"epoch": 3.4,
|
| 1069 |
+
"learning_rate": 3.298076923076923e-05,
|
| 1070 |
+
"loss": 2.9056,
|
| 1071 |
+
"step": 1062
|
| 1072 |
+
},
|
| 1073 |
+
{
|
| 1074 |
+
"epoch": 3.42,
|
| 1075 |
+
"learning_rate": 3.288461538461539e-05,
|
| 1076 |
+
"loss": 2.749,
|
| 1077 |
+
"step": 1068
|
| 1078 |
+
},
|
| 1079 |
+
{
|
| 1080 |
+
"epoch": 3.44,
|
| 1081 |
+
"learning_rate": 3.278846153846154e-05,
|
| 1082 |
+
"loss": 2.7579,
|
| 1083 |
+
"step": 1074
|
| 1084 |
+
},
|
| 1085 |
+
{
|
| 1086 |
+
"epoch": 3.46,
|
| 1087 |
+
"learning_rate": 3.269230769230769e-05,
|
| 1088 |
+
"loss": 2.7444,
|
| 1089 |
+
"step": 1080
|
| 1090 |
+
},
|
| 1091 |
+
{
|
| 1092 |
+
"epoch": 3.48,
|
| 1093 |
+
"learning_rate": 3.2596153846153845e-05,
|
| 1094 |
+
"loss": 2.818,
|
| 1095 |
+
"step": 1086
|
| 1096 |
+
},
|
| 1097 |
+
{
|
| 1098 |
+
"epoch": 3.5,
|
| 1099 |
+
"learning_rate": 3.2500000000000004e-05,
|
| 1100 |
+
"loss": 2.6285,
|
| 1101 |
+
"step": 1092
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 3.52,
|
| 1105 |
+
"learning_rate": 3.2403846153846156e-05,
|
| 1106 |
+
"loss": 2.6473,
|
| 1107 |
+
"step": 1098
|
| 1108 |
+
},
|
| 1109 |
+
{
|
| 1110 |
+
"epoch": 3.54,
|
| 1111 |
+
"learning_rate": 3.230769230769231e-05,
|
| 1112 |
+
"loss": 2.7093,
|
| 1113 |
+
"step": 1104
|
| 1114 |
+
},
|
| 1115 |
+
{
|
| 1116 |
+
"epoch": 3.56,
|
| 1117 |
+
"learning_rate": 3.221153846153847e-05,
|
| 1118 |
+
"loss": 2.732,
|
| 1119 |
+
"step": 1110
|
| 1120 |
+
},
|
| 1121 |
+
{
|
| 1122 |
+
"epoch": 3.58,
|
| 1123 |
+
"learning_rate": 3.211538461538462e-05,
|
| 1124 |
+
"loss": 2.7064,
|
| 1125 |
+
"step": 1116
|
| 1126 |
+
},
|
| 1127 |
+
{
|
| 1128 |
+
"epoch": 3.6,
|
| 1129 |
+
"learning_rate": 3.201923076923077e-05,
|
| 1130 |
+
"loss": 2.9072,
|
| 1131 |
+
"step": 1122
|
| 1132 |
+
},
|
| 1133 |
+
{
|
| 1134 |
+
"epoch": 3.62,
|
| 1135 |
+
"learning_rate": 3.192307692307692e-05,
|
| 1136 |
+
"loss": 2.7544,
|
| 1137 |
+
"step": 1128
|
| 1138 |
+
},
|
| 1139 |
+
{
|
| 1140 |
+
"epoch": 3.63,
|
| 1141 |
+
"learning_rate": 3.1826923076923075e-05,
|
| 1142 |
+
"loss": 2.738,
|
| 1143 |
+
"step": 1134
|
| 1144 |
+
},
|
| 1145 |
+
{
|
| 1146 |
+
"epoch": 3.65,
|
| 1147 |
+
"learning_rate": 3.1730769230769234e-05,
|
| 1148 |
+
"loss": 2.6721,
|
| 1149 |
+
"step": 1140
|
| 1150 |
+
},
|
| 1151 |
+
{
|
| 1152 |
+
"epoch": 3.67,
|
| 1153 |
+
"learning_rate": 3.1634615384615386e-05,
|
| 1154 |
+
"loss": 2.8218,
|
| 1155 |
+
"step": 1146
|
| 1156 |
+
},
|
| 1157 |
+
{
|
| 1158 |
+
"epoch": 3.69,
|
| 1159 |
+
"learning_rate": 3.153846153846154e-05,
|
| 1160 |
+
"loss": 2.7553,
|
| 1161 |
+
"step": 1152
|
| 1162 |
+
},
|
| 1163 |
+
{
|
| 1164 |
+
"epoch": 3.71,
|
| 1165 |
+
"learning_rate": 3.144230769230769e-05,
|
| 1166 |
+
"loss": 2.8186,
|
| 1167 |
+
"step": 1158
|
| 1168 |
+
},
|
| 1169 |
+
{
|
| 1170 |
+
"epoch": 3.73,
|
| 1171 |
+
"learning_rate": 3.134615384615385e-05,
|
| 1172 |
+
"loss": 2.7511,
|
| 1173 |
+
"step": 1164
|
| 1174 |
+
},
|
| 1175 |
+
{
|
| 1176 |
+
"epoch": 3.75,
|
| 1177 |
+
"learning_rate": 3.125e-05,
|
| 1178 |
+
"loss": 2.6013,
|
| 1179 |
+
"step": 1170
|
| 1180 |
+
},
|
| 1181 |
+
{
|
| 1182 |
+
"epoch": 3.77,
|
| 1183 |
+
"learning_rate": 3.115384615384615e-05,
|
| 1184 |
+
"loss": 2.7191,
|
| 1185 |
+
"step": 1176
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 3.79,
|
| 1189 |
+
"learning_rate": 3.105769230769231e-05,
|
| 1190 |
+
"loss": 2.7695,
|
| 1191 |
+
"step": 1182
|
| 1192 |
+
},
|
| 1193 |
+
{
|
| 1194 |
+
"epoch": 3.81,
|
| 1195 |
+
"learning_rate": 3.0961538461538464e-05,
|
| 1196 |
+
"loss": 2.738,
|
| 1197 |
+
"step": 1188
|
| 1198 |
+
},
|
| 1199 |
+
{
|
| 1200 |
+
"epoch": 3.83,
|
| 1201 |
+
"learning_rate": 3.0865384615384616e-05,
|
| 1202 |
+
"loss": 2.7202,
|
| 1203 |
+
"step": 1194
|
| 1204 |
+
},
|
| 1205 |
+
{
|
| 1206 |
+
"epoch": 3.85,
|
| 1207 |
+
"learning_rate": 3.0769230769230774e-05,
|
| 1208 |
+
"loss": 2.6429,
|
| 1209 |
+
"step": 1200
|
| 1210 |
+
},
|
| 1211 |
+
{
|
| 1212 |
+
"epoch": 3.87,
|
| 1213 |
+
"learning_rate": 3.0673076923076926e-05,
|
| 1214 |
+
"loss": 2.6742,
|
| 1215 |
+
"step": 1206
|
| 1216 |
+
},
|
| 1217 |
+
{
|
| 1218 |
+
"epoch": 3.88,
|
| 1219 |
+
"learning_rate": 3.057692307692308e-05,
|
| 1220 |
+
"loss": 2.7048,
|
| 1221 |
+
"step": 1212
|
| 1222 |
+
},
|
| 1223 |
+
{
|
| 1224 |
+
"epoch": 3.9,
|
| 1225 |
+
"learning_rate": 3.0480769230769234e-05,
|
| 1226 |
+
"loss": 2.7016,
|
| 1227 |
+
"step": 1218
|
| 1228 |
+
},
|
| 1229 |
+
{
|
| 1230 |
+
"epoch": 3.92,
|
| 1231 |
+
"learning_rate": 3.0384615384615382e-05,
|
| 1232 |
+
"loss": 2.7375,
|
| 1233 |
+
"step": 1224
|
| 1234 |
+
},
|
| 1235 |
+
{
|
| 1236 |
+
"epoch": 3.94,
|
| 1237 |
+
"learning_rate": 3.0288461538461538e-05,
|
| 1238 |
+
"loss": 2.7926,
|
| 1239 |
+
"step": 1230
|
| 1240 |
+
},
|
| 1241 |
+
{
|
| 1242 |
+
"epoch": 3.96,
|
| 1243 |
+
"learning_rate": 3.0192307692307693e-05,
|
| 1244 |
+
"loss": 2.7038,
|
| 1245 |
+
"step": 1236
|
| 1246 |
+
},
|
| 1247 |
+
{
|
| 1248 |
+
"epoch": 3.98,
|
| 1249 |
+
"learning_rate": 3.0096153846153845e-05,
|
| 1250 |
+
"loss": 2.8294,
|
| 1251 |
+
"step": 1242
|
| 1252 |
+
},
|
| 1253 |
+
{
|
| 1254 |
+
"epoch": 4.0,
|
| 1255 |
+
"learning_rate": 3e-05,
|
| 1256 |
+
"loss": 2.7543,
|
| 1257 |
+
"step": 1248
|
| 1258 |
+
},
|
| 1259 |
+
{
|
| 1260 |
+
"epoch": 4.02,
|
| 1261 |
+
"learning_rate": 2.9903846153846156e-05,
|
| 1262 |
+
"loss": 2.6022,
|
| 1263 |
+
"step": 1254
|
| 1264 |
+
},
|
| 1265 |
+
{
|
| 1266 |
+
"epoch": 4.04,
|
| 1267 |
+
"learning_rate": 2.9807692307692308e-05,
|
| 1268 |
+
"loss": 2.6609,
|
| 1269 |
+
"step": 1260
|
| 1270 |
+
},
|
| 1271 |
+
{
|
| 1272 |
+
"epoch": 4.06,
|
| 1273 |
+
"learning_rate": 2.9711538461538464e-05,
|
| 1274 |
+
"loss": 2.6091,
|
| 1275 |
+
"step": 1266
|
| 1276 |
+
},
|
| 1277 |
+
{
|
| 1278 |
+
"epoch": 4.08,
|
| 1279 |
+
"learning_rate": 2.9615384615384616e-05,
|
| 1280 |
+
"loss": 2.5386,
|
| 1281 |
+
"step": 1272
|
| 1282 |
+
},
|
| 1283 |
+
{
|
| 1284 |
+
"epoch": 4.1,
|
| 1285 |
+
"learning_rate": 2.951923076923077e-05,
|
| 1286 |
+
"loss": 2.6655,
|
| 1287 |
+
"step": 1278
|
| 1288 |
+
},
|
| 1289 |
+
{
|
| 1290 |
+
"epoch": 4.12,
|
| 1291 |
+
"learning_rate": 2.9423076923076926e-05,
|
| 1292 |
+
"loss": 2.6023,
|
| 1293 |
+
"step": 1284
|
| 1294 |
+
},
|
| 1295 |
+
{
|
| 1296 |
+
"epoch": 4.13,
|
| 1297 |
+
"learning_rate": 2.932692307692308e-05,
|
| 1298 |
+
"loss": 2.6491,
|
| 1299 |
+
"step": 1290
|
| 1300 |
+
},
|
| 1301 |
+
{
|
| 1302 |
+
"epoch": 4.15,
|
| 1303 |
+
"learning_rate": 2.9230769230769234e-05,
|
| 1304 |
+
"loss": 2.6548,
|
| 1305 |
+
"step": 1296
|
| 1306 |
+
},
|
| 1307 |
+
{
|
| 1308 |
+
"epoch": 4.17,
|
| 1309 |
+
"learning_rate": 2.913461538461539e-05,
|
| 1310 |
+
"loss": 2.584,
|
| 1311 |
+
"step": 1302
|
| 1312 |
+
},
|
| 1313 |
+
{
|
| 1314 |
+
"epoch": 4.19,
|
| 1315 |
+
"learning_rate": 2.903846153846154e-05,
|
| 1316 |
+
"loss": 2.5447,
|
| 1317 |
+
"step": 1308
|
| 1318 |
+
},
|
| 1319 |
+
{
|
| 1320 |
+
"epoch": 4.21,
|
| 1321 |
+
"learning_rate": 2.8942307692307697e-05,
|
| 1322 |
+
"loss": 2.5931,
|
| 1323 |
+
"step": 1314
|
| 1324 |
+
},
|
| 1325 |
+
{
|
| 1326 |
+
"epoch": 4.23,
|
| 1327 |
+
"learning_rate": 2.8846153846153845e-05,
|
| 1328 |
+
"loss": 2.503,
|
| 1329 |
+
"step": 1320
|
| 1330 |
+
},
|
| 1331 |
+
{
|
| 1332 |
+
"epoch": 4.25,
|
| 1333 |
+
"learning_rate": 2.8749999999999997e-05,
|
| 1334 |
+
"loss": 2.6097,
|
| 1335 |
+
"step": 1326
|
| 1336 |
+
},
|
| 1337 |
+
{
|
| 1338 |
+
"epoch": 4.27,
|
| 1339 |
+
"learning_rate": 2.8653846153846153e-05,
|
| 1340 |
+
"loss": 2.6926,
|
| 1341 |
+
"step": 1332
|
| 1342 |
+
},
|
| 1343 |
+
{
|
| 1344 |
+
"epoch": 4.29,
|
| 1345 |
+
"learning_rate": 2.855769230769231e-05,
|
| 1346 |
+
"loss": 2.5829,
|
| 1347 |
+
"step": 1338
|
| 1348 |
+
},
|
| 1349 |
+
{
|
| 1350 |
+
"epoch": 4.31,
|
| 1351 |
+
"learning_rate": 2.846153846153846e-05,
|
| 1352 |
+
"loss": 2.6712,
|
| 1353 |
+
"step": 1344
|
| 1354 |
+
},
|
| 1355 |
+
{
|
| 1356 |
+
"epoch": 4.33,
|
| 1357 |
+
"learning_rate": 2.8365384615384616e-05,
|
| 1358 |
+
"loss": 2.3988,
|
| 1359 |
+
"step": 1350
|
| 1360 |
+
},
|
| 1361 |
+
{
|
| 1362 |
+
"epoch": 4.35,
|
| 1363 |
+
"learning_rate": 2.826923076923077e-05,
|
| 1364 |
+
"loss": 2.5714,
|
| 1365 |
+
"step": 1356
|
| 1366 |
+
},
|
| 1367 |
+
{
|
| 1368 |
+
"epoch": 4.37,
|
| 1369 |
+
"learning_rate": 2.8173076923076923e-05,
|
| 1370 |
+
"loss": 2.6627,
|
| 1371 |
+
"step": 1362
|
| 1372 |
+
},
|
| 1373 |
+
{
|
| 1374 |
+
"epoch": 4.38,
|
| 1375 |
+
"learning_rate": 2.807692307692308e-05,
|
| 1376 |
+
"loss": 2.7195,
|
| 1377 |
+
"step": 1368
|
| 1378 |
+
},
|
| 1379 |
+
{
|
| 1380 |
+
"epoch": 4.4,
|
| 1381 |
+
"learning_rate": 2.7980769230769234e-05,
|
| 1382 |
+
"loss": 2.7181,
|
| 1383 |
+
"step": 1374
|
| 1384 |
+
},
|
| 1385 |
+
{
|
| 1386 |
+
"epoch": 4.42,
|
| 1387 |
+
"learning_rate": 2.7884615384615386e-05,
|
| 1388 |
+
"loss": 2.626,
|
| 1389 |
+
"step": 1380
|
| 1390 |
+
},
|
| 1391 |
+
{
|
| 1392 |
+
"epoch": 4.44,
|
| 1393 |
+
"learning_rate": 2.778846153846154e-05,
|
| 1394 |
+
"loss": 2.638,
|
| 1395 |
+
"step": 1386
|
| 1396 |
+
},
|
| 1397 |
+
{
|
| 1398 |
+
"epoch": 4.46,
|
| 1399 |
+
"learning_rate": 2.7692307692307694e-05,
|
| 1400 |
+
"loss": 2.6275,
|
| 1401 |
+
"step": 1392
|
| 1402 |
+
},
|
| 1403 |
+
{
|
| 1404 |
+
"epoch": 4.48,
|
| 1405 |
+
"learning_rate": 2.759615384615385e-05,
|
| 1406 |
+
"loss": 2.6399,
|
| 1407 |
+
"step": 1398
|
| 1408 |
+
},
|
| 1409 |
+
{
|
| 1410 |
+
"epoch": 4.5,
|
| 1411 |
+
"learning_rate": 2.7500000000000004e-05,
|
| 1412 |
+
"loss": 2.6894,
|
| 1413 |
+
"step": 1404
|
| 1414 |
+
},
|
| 1415 |
+
{
|
| 1416 |
+
"epoch": 4.52,
|
| 1417 |
+
"learning_rate": 2.7403846153846156e-05,
|
| 1418 |
+
"loss": 2.5917,
|
| 1419 |
+
"step": 1410
|
| 1420 |
+
},
|
| 1421 |
+
{
|
| 1422 |
+
"epoch": 4.54,
|
| 1423 |
+
"learning_rate": 2.7307692307692305e-05,
|
| 1424 |
+
"loss": 2.5564,
|
| 1425 |
+
"step": 1416
|
| 1426 |
+
},
|
| 1427 |
+
{
|
| 1428 |
+
"epoch": 4.56,
|
| 1429 |
+
"learning_rate": 2.721153846153846e-05,
|
| 1430 |
+
"loss": 2.6134,
|
| 1431 |
+
"step": 1422
|
| 1432 |
+
},
|
| 1433 |
+
{
|
| 1434 |
+
"epoch": 4.58,
|
| 1435 |
+
"learning_rate": 2.7115384615384616e-05,
|
| 1436 |
+
"loss": 2.6375,
|
| 1437 |
+
"step": 1428
|
| 1438 |
+
},
|
| 1439 |
+
{
|
| 1440 |
+
"epoch": 4.6,
|
| 1441 |
+
"learning_rate": 2.7019230769230768e-05,
|
| 1442 |
+
"loss": 2.6301,
|
| 1443 |
+
"step": 1434
|
| 1444 |
+
},
|
| 1445 |
+
{
|
| 1446 |
+
"epoch": 4.62,
|
| 1447 |
+
"learning_rate": 2.6923076923076923e-05,
|
| 1448 |
+
"loss": 2.5679,
|
| 1449 |
+
"step": 1440
|
| 1450 |
+
},
|
| 1451 |
+
{
|
| 1452 |
+
"epoch": 4.63,
|
| 1453 |
+
"learning_rate": 2.682692307692308e-05,
|
| 1454 |
+
"loss": 2.6704,
|
| 1455 |
+
"step": 1446
|
| 1456 |
+
},
|
| 1457 |
+
{
|
| 1458 |
+
"epoch": 4.65,
|
| 1459 |
+
"learning_rate": 2.673076923076923e-05,
|
| 1460 |
+
"loss": 2.6778,
|
| 1461 |
+
"step": 1452
|
| 1462 |
+
},
|
| 1463 |
+
{
|
| 1464 |
+
"epoch": 4.67,
|
| 1465 |
+
"learning_rate": 2.6634615384615386e-05,
|
| 1466 |
+
"loss": 2.5626,
|
| 1467 |
+
"step": 1458
|
| 1468 |
+
},
|
| 1469 |
+
{
|
| 1470 |
+
"epoch": 4.69,
|
| 1471 |
+
"learning_rate": 2.6538461538461538e-05,
|
| 1472 |
+
"loss": 2.5557,
|
| 1473 |
+
"step": 1464
|
| 1474 |
+
},
|
| 1475 |
+
{
|
| 1476 |
+
"epoch": 4.71,
|
| 1477 |
+
"learning_rate": 2.6442307692307694e-05,
|
| 1478 |
+
"loss": 2.5245,
|
| 1479 |
+
"step": 1470
|
| 1480 |
+
},
|
| 1481 |
+
{
|
| 1482 |
+
"epoch": 4.73,
|
| 1483 |
+
"learning_rate": 2.634615384615385e-05,
|
| 1484 |
+
"loss": 2.5189,
|
| 1485 |
+
"step": 1476
|
| 1486 |
+
},
|
| 1487 |
+
{
|
| 1488 |
+
"epoch": 4.75,
|
| 1489 |
+
"learning_rate": 2.625e-05,
|
| 1490 |
+
"loss": 2.4994,
|
| 1491 |
+
"step": 1482
|
| 1492 |
+
},
|
| 1493 |
+
{
|
| 1494 |
+
"epoch": 4.77,
|
| 1495 |
+
"learning_rate": 2.6153846153846157e-05,
|
| 1496 |
+
"loss": 2.472,
|
| 1497 |
+
"step": 1488
|
| 1498 |
+
},
|
| 1499 |
+
{
|
| 1500 |
+
"epoch": 4.79,
|
| 1501 |
+
"learning_rate": 2.6057692307692312e-05,
|
| 1502 |
+
"loss": 2.5451,
|
| 1503 |
+
"step": 1494
|
| 1504 |
+
},
|
| 1505 |
+
{
|
| 1506 |
+
"epoch": 4.81,
|
| 1507 |
+
"learning_rate": 2.5961538461538464e-05,
|
| 1508 |
+
"loss": 2.5559,
|
| 1509 |
+
"step": 1500
|
| 1510 |
+
},
|
| 1511 |
+
{
|
| 1512 |
+
"epoch": 4.83,
|
| 1513 |
+
"learning_rate": 2.586538461538462e-05,
|
| 1514 |
+
"loss": 2.5902,
|
| 1515 |
+
"step": 1506
|
| 1516 |
+
},
|
| 1517 |
+
{
|
| 1518 |
+
"epoch": 4.85,
|
| 1519 |
+
"learning_rate": 2.5769230769230768e-05,
|
| 1520 |
+
"loss": 2.6841,
|
| 1521 |
+
"step": 1512
|
| 1522 |
+
},
|
| 1523 |
+
{
|
| 1524 |
+
"epoch": 4.87,
|
| 1525 |
+
"learning_rate": 2.5673076923076923e-05,
|
| 1526 |
+
"loss": 2.6092,
|
| 1527 |
+
"step": 1518
|
| 1528 |
+
},
|
| 1529 |
+
{
|
| 1530 |
+
"epoch": 4.88,
|
| 1531 |
+
"learning_rate": 2.5576923076923075e-05,
|
| 1532 |
+
"loss": 2.5646,
|
| 1533 |
+
"step": 1524
|
| 1534 |
+
},
|
| 1535 |
+
{
|
| 1536 |
+
"epoch": 4.9,
|
| 1537 |
+
"learning_rate": 2.548076923076923e-05,
|
| 1538 |
+
"loss": 2.5201,
|
| 1539 |
+
"step": 1530
|
| 1540 |
+
},
|
| 1541 |
+
{
|
| 1542 |
+
"epoch": 4.92,
|
| 1543 |
+
"learning_rate": 2.5384615384615383e-05,
|
| 1544 |
+
"loss": 2.5622,
|
| 1545 |
+
"step": 1536
|
| 1546 |
+
},
|
| 1547 |
+
{
|
| 1548 |
+
"epoch": 4.94,
|
| 1549 |
+
"learning_rate": 2.528846153846154e-05,
|
| 1550 |
+
"loss": 2.6323,
|
| 1551 |
+
"step": 1542
|
| 1552 |
+
},
|
| 1553 |
+
{
|
| 1554 |
+
"epoch": 4.96,
|
| 1555 |
+
"learning_rate": 2.5192307692307694e-05,
|
| 1556 |
+
"loss": 2.7179,
|
| 1557 |
+
"step": 1548
|
| 1558 |
+
},
|
| 1559 |
+
{
|
| 1560 |
+
"epoch": 4.98,
|
| 1561 |
+
"learning_rate": 2.5096153846153846e-05,
|
| 1562 |
+
"loss": 2.6441,
|
| 1563 |
+
"step": 1554
|
| 1564 |
+
},
|
| 1565 |
+
{
|
| 1566 |
+
"epoch": 5.0,
|
| 1567 |
+
"learning_rate": 2.5e-05,
|
| 1568 |
+
"loss": 2.668,
|
| 1569 |
+
"step": 1560
|
| 1570 |
+
},
|
| 1571 |
+
{
|
| 1572 |
+
"epoch": 5.02,
|
| 1573 |
+
"learning_rate": 2.4903846153846157e-05,
|
| 1574 |
+
"loss": 2.5535,
|
| 1575 |
+
"step": 1566
|
| 1576 |
+
},
|
| 1577 |
+
{
|
| 1578 |
+
"epoch": 5.04,
|
| 1579 |
+
"learning_rate": 2.480769230769231e-05,
|
| 1580 |
+
"loss": 2.4569,
|
| 1581 |
+
"step": 1572
|
| 1582 |
+
},
|
| 1583 |
+
{
|
| 1584 |
+
"epoch": 5.06,
|
| 1585 |
+
"learning_rate": 2.4711538461538464e-05,
|
| 1586 |
+
"loss": 2.5378,
|
| 1587 |
+
"step": 1578
|
| 1588 |
+
},
|
| 1589 |
+
{
|
| 1590 |
+
"epoch": 5.08,
|
| 1591 |
+
"learning_rate": 2.461538461538462e-05,
|
| 1592 |
+
"loss": 2.537,
|
| 1593 |
+
"step": 1584
|
| 1594 |
+
},
|
| 1595 |
+
{
|
| 1596 |
+
"epoch": 5.1,
|
| 1597 |
+
"learning_rate": 2.4519230769230768e-05,
|
| 1598 |
+
"loss": 2.4913,
|
| 1599 |
+
"step": 1590
|
| 1600 |
+
},
|
| 1601 |
+
{
|
| 1602 |
+
"epoch": 5.12,
|
| 1603 |
+
"learning_rate": 2.4423076923076924e-05,
|
| 1604 |
+
"loss": 2.5156,
|
| 1605 |
+
"step": 1596
|
| 1606 |
+
},
|
| 1607 |
+
{
|
| 1608 |
+
"epoch": 5.13,
|
| 1609 |
+
"learning_rate": 2.432692307692308e-05,
|
| 1610 |
+
"loss": 2.493,
|
| 1611 |
+
"step": 1602
|
| 1612 |
+
},
|
| 1613 |
+
{
|
| 1614 |
+
"epoch": 5.15,
|
| 1615 |
+
"learning_rate": 2.423076923076923e-05,
|
| 1616 |
+
"loss": 2.5357,
|
| 1617 |
+
"step": 1608
|
| 1618 |
+
},
|
| 1619 |
+
{
|
| 1620 |
+
"epoch": 5.17,
|
| 1621 |
+
"learning_rate": 2.4134615384615386e-05,
|
| 1622 |
+
"loss": 2.4508,
|
| 1623 |
+
"step": 1614
|
| 1624 |
+
},
|
| 1625 |
+
{
|
| 1626 |
+
"epoch": 5.19,
|
| 1627 |
+
"learning_rate": 2.4038461538461542e-05,
|
| 1628 |
+
"loss": 2.5045,
|
| 1629 |
+
"step": 1620
|
| 1630 |
+
},
|
| 1631 |
+
{
|
| 1632 |
+
"epoch": 5.21,
|
| 1633 |
+
"learning_rate": 2.3942307692307694e-05,
|
| 1634 |
+
"loss": 2.4617,
|
| 1635 |
+
"step": 1626
|
| 1636 |
+
},
|
| 1637 |
+
{
|
| 1638 |
+
"epoch": 5.23,
|
| 1639 |
+
"learning_rate": 2.384615384615385e-05,
|
| 1640 |
+
"loss": 2.5392,
|
| 1641 |
+
"step": 1632
|
| 1642 |
+
},
|
| 1643 |
+
{
|
| 1644 |
+
"epoch": 5.25,
|
| 1645 |
+
"learning_rate": 2.375e-05,
|
| 1646 |
+
"loss": 2.3976,
|
| 1647 |
+
"step": 1638
|
| 1648 |
+
},
|
| 1649 |
+
{
|
| 1650 |
+
"epoch": 5.27,
|
| 1651 |
+
"learning_rate": 2.3653846153846153e-05,
|
| 1652 |
+
"loss": 2.5288,
|
| 1653 |
+
"step": 1644
|
| 1654 |
+
},
|
| 1655 |
+
{
|
| 1656 |
+
"epoch": 5.29,
|
| 1657 |
+
"learning_rate": 2.355769230769231e-05,
|
| 1658 |
+
"loss": 2.5243,
|
| 1659 |
+
"step": 1650
|
| 1660 |
+
},
|
| 1661 |
+
{
|
| 1662 |
+
"epoch": 5.31,
|
| 1663 |
+
"learning_rate": 2.3461538461538464e-05,
|
| 1664 |
+
"loss": 2.5018,
|
| 1665 |
+
"step": 1656
|
| 1666 |
+
},
|
| 1667 |
+
{
|
| 1668 |
+
"epoch": 5.33,
|
| 1669 |
+
"learning_rate": 2.3365384615384616e-05,
|
| 1670 |
+
"loss": 2.5947,
|
| 1671 |
+
"step": 1662
|
| 1672 |
+
},
|
| 1673 |
+
{
|
| 1674 |
+
"epoch": 5.35,
|
| 1675 |
+
"learning_rate": 2.326923076923077e-05,
|
| 1676 |
+
"loss": 2.5226,
|
| 1677 |
+
"step": 1668
|
| 1678 |
+
},
|
| 1679 |
+
{
|
| 1680 |
+
"epoch": 5.37,
|
| 1681 |
+
"learning_rate": 2.3173076923076924e-05,
|
| 1682 |
+
"loss": 2.4687,
|
| 1683 |
+
"step": 1674
|
| 1684 |
+
},
|
| 1685 |
+
{
|
| 1686 |
+
"epoch": 5.38,
|
| 1687 |
+
"learning_rate": 2.307692307692308e-05,
|
| 1688 |
+
"loss": 2.5297,
|
| 1689 |
+
"step": 1680
|
| 1690 |
+
},
|
| 1691 |
+
{
|
| 1692 |
+
"epoch": 5.4,
|
| 1693 |
+
"learning_rate": 2.298076923076923e-05,
|
| 1694 |
+
"loss": 2.5457,
|
| 1695 |
+
"step": 1686
|
| 1696 |
+
},
|
| 1697 |
+
{
|
| 1698 |
+
"epoch": 5.42,
|
| 1699 |
+
"learning_rate": 2.2884615384615387e-05,
|
| 1700 |
+
"loss": 2.4218,
|
| 1701 |
+
"step": 1692
|
| 1702 |
+
},
|
| 1703 |
+
{
|
| 1704 |
+
"epoch": 5.44,
|
| 1705 |
+
"learning_rate": 2.278846153846154e-05,
|
| 1706 |
+
"loss": 2.5219,
|
| 1707 |
+
"step": 1698
|
| 1708 |
+
},
|
| 1709 |
+
{
|
| 1710 |
+
"epoch": 5.46,
|
| 1711 |
+
"learning_rate": 2.2692307692307694e-05,
|
| 1712 |
+
"loss": 2.4858,
|
| 1713 |
+
"step": 1704
|
| 1714 |
+
},
|
| 1715 |
+
{
|
| 1716 |
+
"epoch": 5.48,
|
| 1717 |
+
"learning_rate": 2.2596153846153846e-05,
|
| 1718 |
+
"loss": 2.4715,
|
| 1719 |
+
"step": 1710
|
| 1720 |
+
},
|
| 1721 |
+
{
|
| 1722 |
+
"epoch": 5.5,
|
| 1723 |
+
"learning_rate": 2.25e-05,
|
| 1724 |
+
"loss": 2.3294,
|
| 1725 |
+
"step": 1716
|
| 1726 |
+
},
|
| 1727 |
+
{
|
| 1728 |
+
"epoch": 5.52,
|
| 1729 |
+
"learning_rate": 2.2403846153846157e-05,
|
| 1730 |
+
"loss": 2.5062,
|
| 1731 |
+
"step": 1722
|
| 1732 |
+
},
|
| 1733 |
+
{
|
| 1734 |
+
"epoch": 5.54,
|
| 1735 |
+
"learning_rate": 2.230769230769231e-05,
|
| 1736 |
+
"loss": 2.392,
|
| 1737 |
+
"step": 1728
|
| 1738 |
+
},
|
| 1739 |
+
{
|
| 1740 |
+
"epoch": 5.56,
|
| 1741 |
+
"learning_rate": 2.221153846153846e-05,
|
| 1742 |
+
"loss": 2.4092,
|
| 1743 |
+
"step": 1734
|
| 1744 |
+
},
|
| 1745 |
+
{
|
| 1746 |
+
"epoch": 5.58,
|
| 1747 |
+
"learning_rate": 2.2115384615384616e-05,
|
| 1748 |
+
"loss": 2.4257,
|
| 1749 |
+
"step": 1740
|
| 1750 |
+
},
|
| 1751 |
+
{
|
| 1752 |
+
"epoch": 5.6,
|
| 1753 |
+
"learning_rate": 2.201923076923077e-05,
|
| 1754 |
+
"loss": 2.4426,
|
| 1755 |
+
"step": 1746
|
| 1756 |
+
},
|
| 1757 |
+
{
|
| 1758 |
+
"epoch": 5.62,
|
| 1759 |
+
"learning_rate": 2.1923076923076924e-05,
|
| 1760 |
+
"loss": 2.4672,
|
| 1761 |
+
"step": 1752
|
| 1762 |
+
},
|
| 1763 |
+
{
|
| 1764 |
+
"epoch": 5.63,
|
| 1765 |
+
"learning_rate": 2.182692307692308e-05,
|
| 1766 |
+
"loss": 2.5468,
|
| 1767 |
+
"step": 1758
|
| 1768 |
+
},
|
| 1769 |
+
{
|
| 1770 |
+
"epoch": 5.65,
|
| 1771 |
+
"learning_rate": 2.173076923076923e-05,
|
| 1772 |
+
"loss": 2.3917,
|
| 1773 |
+
"step": 1764
|
| 1774 |
+
},
|
| 1775 |
+
{
|
| 1776 |
+
"epoch": 5.67,
|
| 1777 |
+
"learning_rate": 2.1634615384615387e-05,
|
| 1778 |
+
"loss": 2.3895,
|
| 1779 |
+
"step": 1770
|
| 1780 |
+
},
|
| 1781 |
+
{
|
| 1782 |
+
"epoch": 5.69,
|
| 1783 |
+
"learning_rate": 2.1538461538461542e-05,
|
| 1784 |
+
"loss": 2.5246,
|
| 1785 |
+
"step": 1776
|
| 1786 |
+
},
|
| 1787 |
+
{
|
| 1788 |
+
"epoch": 5.71,
|
| 1789 |
+
"learning_rate": 2.144230769230769e-05,
|
| 1790 |
+
"loss": 2.5025,
|
| 1791 |
+
"step": 1782
|
| 1792 |
+
},
|
| 1793 |
+
{
|
| 1794 |
+
"epoch": 5.73,
|
| 1795 |
+
"learning_rate": 2.1346153846153846e-05,
|
| 1796 |
+
"loss": 2.4898,
|
| 1797 |
+
"step": 1788
|
| 1798 |
+
},
|
| 1799 |
+
{
|
| 1800 |
+
"epoch": 5.75,
|
| 1801 |
+
"learning_rate": 2.125e-05,
|
| 1802 |
+
"loss": 2.5358,
|
| 1803 |
+
"step": 1794
|
| 1804 |
+
},
|
| 1805 |
+
{
|
| 1806 |
+
"epoch": 5.77,
|
| 1807 |
+
"learning_rate": 2.1153846153846154e-05,
|
| 1808 |
+
"loss": 2.5346,
|
| 1809 |
+
"step": 1800
|
| 1810 |
+
},
|
| 1811 |
+
{
|
| 1812 |
+
"epoch": 5.79,
|
| 1813 |
+
"learning_rate": 2.105769230769231e-05,
|
| 1814 |
+
"loss": 2.3057,
|
| 1815 |
+
"step": 1806
|
| 1816 |
+
},
|
| 1817 |
+
{
|
| 1818 |
+
"epoch": 5.81,
|
| 1819 |
+
"learning_rate": 2.0961538461538464e-05,
|
| 1820 |
+
"loss": 2.5392,
|
| 1821 |
+
"step": 1812
|
| 1822 |
+
},
|
| 1823 |
+
{
|
| 1824 |
+
"epoch": 5.83,
|
| 1825 |
+
"learning_rate": 2.0865384615384616e-05,
|
| 1826 |
+
"loss": 2.5815,
|
| 1827 |
+
"step": 1818
|
| 1828 |
+
},
|
| 1829 |
+
{
|
| 1830 |
+
"epoch": 5.85,
|
| 1831 |
+
"learning_rate": 2.0769230769230772e-05,
|
| 1832 |
+
"loss": 2.2718,
|
| 1833 |
+
"step": 1824
|
| 1834 |
+
},
|
| 1835 |
+
{
|
| 1836 |
+
"epoch": 5.87,
|
| 1837 |
+
"learning_rate": 2.0673076923076924e-05,
|
| 1838 |
+
"loss": 2.474,
|
| 1839 |
+
"step": 1830
|
| 1840 |
+
},
|
| 1841 |
+
{
|
| 1842 |
+
"epoch": 5.88,
|
| 1843 |
+
"learning_rate": 2.0576923076923076e-05,
|
| 1844 |
+
"loss": 2.4821,
|
| 1845 |
+
"step": 1836
|
| 1846 |
+
},
|
| 1847 |
+
{
|
| 1848 |
+
"epoch": 5.9,
|
| 1849 |
+
"learning_rate": 2.048076923076923e-05,
|
| 1850 |
+
"loss": 2.5334,
|
| 1851 |
+
"step": 1842
|
| 1852 |
+
},
|
| 1853 |
+
{
|
| 1854 |
+
"epoch": 5.92,
|
| 1855 |
+
"learning_rate": 2.0384615384615387e-05,
|
| 1856 |
+
"loss": 2.3877,
|
| 1857 |
+
"step": 1848
|
| 1858 |
+
},
|
| 1859 |
+
{
|
| 1860 |
+
"epoch": 5.94,
|
| 1861 |
+
"learning_rate": 2.028846153846154e-05,
|
| 1862 |
+
"loss": 2.3301,
|
| 1863 |
+
"step": 1854
|
| 1864 |
+
},
|
| 1865 |
+
{
|
| 1866 |
+
"epoch": 5.96,
|
| 1867 |
+
"learning_rate": 2.0192307692307694e-05,
|
| 1868 |
+
"loss": 2.545,
|
| 1869 |
+
"step": 1860
|
| 1870 |
+
},
|
| 1871 |
+
{
|
| 1872 |
+
"epoch": 5.98,
|
| 1873 |
+
"learning_rate": 2.0096153846153846e-05,
|
| 1874 |
+
"loss": 2.3865,
|
| 1875 |
+
"step": 1866
|
| 1876 |
+
},
|
| 1877 |
+
{
|
| 1878 |
+
"epoch": 6.0,
|
| 1879 |
+
"learning_rate": 2e-05,
|
| 1880 |
+
"loss": 2.4531,
|
| 1881 |
+
"step": 1872
|
| 1882 |
+
},
|
| 1883 |
+
{
|
| 1884 |
+
"epoch": 6.02,
|
| 1885 |
+
"learning_rate": 1.9903846153846154e-05,
|
| 1886 |
+
"loss": 2.4046,
|
| 1887 |
+
"step": 1878
|
| 1888 |
+
},
|
| 1889 |
+
{
|
| 1890 |
+
"epoch": 6.04,
|
| 1891 |
+
"learning_rate": 1.980769230769231e-05,
|
| 1892 |
+
"loss": 2.476,
|
| 1893 |
+
"step": 1884
|
| 1894 |
+
},
|
| 1895 |
+
{
|
| 1896 |
+
"epoch": 6.06,
|
| 1897 |
+
"learning_rate": 1.971153846153846e-05,
|
| 1898 |
+
"loss": 2.3547,
|
| 1899 |
+
"step": 1890
|
| 1900 |
+
},
|
| 1901 |
+
{
|
| 1902 |
+
"epoch": 6.08,
|
| 1903 |
+
"learning_rate": 1.9615384615384617e-05,
|
| 1904 |
+
"loss": 2.5396,
|
| 1905 |
+
"step": 1896
|
| 1906 |
+
},
|
| 1907 |
+
{
|
| 1908 |
+
"epoch": 6.1,
|
| 1909 |
+
"learning_rate": 1.951923076923077e-05,
|
| 1910 |
+
"loss": 2.3868,
|
| 1911 |
+
"step": 1902
|
| 1912 |
+
},
|
| 1913 |
+
{
|
| 1914 |
+
"epoch": 6.12,
|
| 1915 |
+
"learning_rate": 1.9423076923076924e-05,
|
| 1916 |
+
"loss": 2.2733,
|
| 1917 |
+
"step": 1908
|
| 1918 |
+
},
|
| 1919 |
+
{
|
| 1920 |
+
"epoch": 6.13,
|
| 1921 |
+
"learning_rate": 1.932692307692308e-05,
|
| 1922 |
+
"loss": 2.283,
|
| 1923 |
+
"step": 1914
|
| 1924 |
+
},
|
| 1925 |
+
{
|
| 1926 |
+
"epoch": 6.15,
|
| 1927 |
+
"learning_rate": 1.923076923076923e-05,
|
| 1928 |
+
"loss": 2.4231,
|
| 1929 |
+
"step": 1920
|
| 1930 |
+
},
|
| 1931 |
+
{
|
| 1932 |
+
"epoch": 6.17,
|
| 1933 |
+
"learning_rate": 1.9134615384615383e-05,
|
| 1934 |
+
"loss": 2.375,
|
| 1935 |
+
"step": 1926
|
| 1936 |
+
},
|
| 1937 |
+
{
|
| 1938 |
+
"epoch": 6.19,
|
| 1939 |
+
"learning_rate": 1.903846153846154e-05,
|
| 1940 |
+
"loss": 2.3824,
|
| 1941 |
+
"step": 1932
|
| 1942 |
+
},
|
| 1943 |
+
{
|
| 1944 |
+
"epoch": 6.21,
|
| 1945 |
+
"learning_rate": 1.894230769230769e-05,
|
| 1946 |
+
"loss": 2.3903,
|
| 1947 |
+
"step": 1938
|
| 1948 |
+
},
|
| 1949 |
+
{
|
| 1950 |
+
"epoch": 6.23,
|
| 1951 |
+
"learning_rate": 1.8846153846153846e-05,
|
| 1952 |
+
"loss": 2.2854,
|
| 1953 |
+
"step": 1944
|
| 1954 |
+
},
|
| 1955 |
+
{
|
| 1956 |
+
"epoch": 6.25,
|
| 1957 |
+
"learning_rate": 1.8750000000000002e-05,
|
| 1958 |
+
"loss": 2.2358,
|
| 1959 |
+
"step": 1950
|
| 1960 |
+
},
|
| 1961 |
+
{
|
| 1962 |
+
"epoch": 6.27,
|
| 1963 |
+
"learning_rate": 1.8653846153846154e-05,
|
| 1964 |
+
"loss": 2.4751,
|
| 1965 |
+
"step": 1956
|
| 1966 |
+
},
|
| 1967 |
+
{
|
| 1968 |
+
"epoch": 6.29,
|
| 1969 |
+
"learning_rate": 1.855769230769231e-05,
|
| 1970 |
+
"loss": 2.3434,
|
| 1971 |
+
"step": 1962
|
| 1972 |
+
},
|
| 1973 |
+
{
|
| 1974 |
+
"epoch": 6.31,
|
| 1975 |
+
"learning_rate": 1.8461538461538465e-05,
|
| 1976 |
+
"loss": 2.4088,
|
| 1977 |
+
"step": 1968
|
| 1978 |
+
},
|
| 1979 |
+
{
|
| 1980 |
+
"epoch": 6.33,
|
| 1981 |
+
"learning_rate": 1.8365384615384617e-05,
|
| 1982 |
+
"loss": 2.2746,
|
| 1983 |
+
"step": 1974
|
| 1984 |
+
},
|
| 1985 |
+
{
|
| 1986 |
+
"epoch": 6.35,
|
| 1987 |
+
"learning_rate": 1.826923076923077e-05,
|
| 1988 |
+
"loss": 2.4396,
|
| 1989 |
+
"step": 1980
|
| 1990 |
+
},
|
| 1991 |
+
{
|
| 1992 |
+
"epoch": 6.37,
|
| 1993 |
+
"learning_rate": 1.8173076923076924e-05,
|
| 1994 |
+
"loss": 2.3663,
|
| 1995 |
+
"step": 1986
|
| 1996 |
+
},
|
| 1997 |
+
{
|
| 1998 |
+
"epoch": 6.38,
|
| 1999 |
+
"learning_rate": 1.8076923076923076e-05,
|
| 2000 |
+
"loss": 2.4488,
|
| 2001 |
+
"step": 1992
|
| 2002 |
+
},
|
| 2003 |
+
{
|
| 2004 |
+
"epoch": 6.4,
|
| 2005 |
+
"learning_rate": 1.798076923076923e-05,
|
| 2006 |
+
"loss": 2.3495,
|
| 2007 |
+
"step": 1998
|
| 2008 |
+
},
|
| 2009 |
+
{
|
| 2010 |
+
"epoch": 6.42,
|
| 2011 |
+
"learning_rate": 1.7884615384615387e-05,
|
| 2012 |
+
"loss": 2.4099,
|
| 2013 |
+
"step": 2004
|
| 2014 |
+
},
|
| 2015 |
+
{
|
| 2016 |
+
"epoch": 6.44,
|
| 2017 |
+
"learning_rate": 1.778846153846154e-05,
|
| 2018 |
+
"loss": 2.4343,
|
| 2019 |
+
"step": 2010
|
| 2020 |
+
},
|
| 2021 |
+
{
|
| 2022 |
+
"epoch": 6.46,
|
| 2023 |
+
"learning_rate": 1.7692307692307694e-05,
|
| 2024 |
+
"loss": 2.3508,
|
| 2025 |
+
"step": 2016
|
| 2026 |
+
},
|
| 2027 |
+
{
|
| 2028 |
+
"epoch": 6.48,
|
| 2029 |
+
"learning_rate": 1.759615384615385e-05,
|
| 2030 |
+
"loss": 2.4724,
|
| 2031 |
+
"step": 2022
|
| 2032 |
+
},
|
| 2033 |
+
{
|
| 2034 |
+
"epoch": 6.5,
|
| 2035 |
+
"learning_rate": 1.75e-05,
|
| 2036 |
+
"loss": 2.4746,
|
| 2037 |
+
"step": 2028
|
| 2038 |
+
},
|
| 2039 |
+
{
|
| 2040 |
+
"epoch": 6.52,
|
| 2041 |
+
"learning_rate": 1.7403846153846154e-05,
|
| 2042 |
+
"loss": 2.3899,
|
| 2043 |
+
"step": 2034
|
| 2044 |
+
},
|
| 2045 |
+
{
|
| 2046 |
+
"epoch": 6.54,
|
| 2047 |
+
"learning_rate": 1.730769230769231e-05,
|
| 2048 |
+
"loss": 2.3465,
|
| 2049 |
+
"step": 2040
|
| 2050 |
+
},
|
| 2051 |
+
{
|
| 2052 |
+
"epoch": 6.56,
|
| 2053 |
+
"learning_rate": 1.721153846153846e-05,
|
| 2054 |
+
"loss": 2.3984,
|
| 2055 |
+
"step": 2046
|
| 2056 |
+
},
|
| 2057 |
+
{
|
| 2058 |
+
"epoch": 6.58,
|
| 2059 |
+
"learning_rate": 1.7115384615384617e-05,
|
| 2060 |
+
"loss": 2.3041,
|
| 2061 |
+
"step": 2052
|
| 2062 |
+
},
|
| 2063 |
+
{
|
| 2064 |
+
"epoch": 6.6,
|
| 2065 |
+
"learning_rate": 1.7019230769230772e-05,
|
| 2066 |
+
"loss": 2.3035,
|
| 2067 |
+
"step": 2058
|
| 2068 |
+
},
|
| 2069 |
+
{
|
| 2070 |
+
"epoch": 6.62,
|
| 2071 |
+
"learning_rate": 1.6923076923076924e-05,
|
| 2072 |
+
"loss": 2.2846,
|
| 2073 |
+
"step": 2064
|
| 2074 |
+
},
|
| 2075 |
+
{
|
| 2076 |
+
"epoch": 6.63,
|
| 2077 |
+
"learning_rate": 1.682692307692308e-05,
|
| 2078 |
+
"loss": 2.4577,
|
| 2079 |
+
"step": 2070
|
| 2080 |
+
},
|
| 2081 |
+
{
|
| 2082 |
+
"epoch": 6.65,
|
| 2083 |
+
"learning_rate": 1.673076923076923e-05,
|
| 2084 |
+
"loss": 2.3338,
|
| 2085 |
+
"step": 2076
|
| 2086 |
+
},
|
| 2087 |
+
{
|
| 2088 |
+
"epoch": 6.67,
|
| 2089 |
+
"learning_rate": 1.6634615384615384e-05,
|
| 2090 |
+
"loss": 2.3623,
|
| 2091 |
+
"step": 2082
|
| 2092 |
+
},
|
| 2093 |
+
{
|
| 2094 |
+
"epoch": 6.69,
|
| 2095 |
+
"learning_rate": 1.653846153846154e-05,
|
| 2096 |
+
"loss": 2.3601,
|
| 2097 |
+
"step": 2088
|
| 2098 |
+
},
|
| 2099 |
+
{
|
| 2100 |
+
"epoch": 6.71,
|
| 2101 |
+
"learning_rate": 1.6442307692307695e-05,
|
| 2102 |
+
"loss": 2.34,
|
| 2103 |
+
"step": 2094
|
| 2104 |
+
},
|
| 2105 |
+
{
|
| 2106 |
+
"epoch": 6.73,
|
| 2107 |
+
"learning_rate": 1.6346153846153847e-05,
|
| 2108 |
+
"loss": 2.3451,
|
| 2109 |
+
"step": 2100
|
| 2110 |
+
},
|
| 2111 |
+
{
|
| 2112 |
+
"epoch": 6.75,
|
| 2113 |
+
"learning_rate": 1.6250000000000002e-05,
|
| 2114 |
+
"loss": 2.3392,
|
| 2115 |
+
"step": 2106
|
| 2116 |
+
},
|
| 2117 |
+
{
|
| 2118 |
+
"epoch": 6.77,
|
| 2119 |
+
"learning_rate": 1.6153846153846154e-05,
|
| 2120 |
+
"loss": 2.4011,
|
| 2121 |
+
"step": 2112
|
| 2122 |
+
},
|
| 2123 |
+
{
|
| 2124 |
+
"epoch": 6.79,
|
| 2125 |
+
"learning_rate": 1.605769230769231e-05,
|
| 2126 |
+
"loss": 2.382,
|
| 2127 |
+
"step": 2118
|
| 2128 |
+
},
|
| 2129 |
+
{
|
| 2130 |
+
"epoch": 6.81,
|
| 2131 |
+
"learning_rate": 1.596153846153846e-05,
|
| 2132 |
+
"loss": 2.4005,
|
| 2133 |
+
"step": 2124
|
| 2134 |
+
},
|
| 2135 |
+
{
|
| 2136 |
+
"epoch": 6.83,
|
| 2137 |
+
"learning_rate": 1.5865384615384617e-05,
|
| 2138 |
+
"loss": 2.3398,
|
| 2139 |
+
"step": 2130
|
| 2140 |
+
},
|
| 2141 |
+
{
|
| 2142 |
+
"epoch": 6.85,
|
| 2143 |
+
"learning_rate": 1.576923076923077e-05,
|
| 2144 |
+
"loss": 2.3156,
|
| 2145 |
+
"step": 2136
|
| 2146 |
+
},
|
| 2147 |
+
{
|
| 2148 |
+
"epoch": 6.87,
|
| 2149 |
+
"learning_rate": 1.5673076923076924e-05,
|
| 2150 |
+
"loss": 2.3249,
|
| 2151 |
+
"step": 2142
|
| 2152 |
+
},
|
| 2153 |
+
{
|
| 2154 |
+
"epoch": 6.88,
|
| 2155 |
+
"learning_rate": 1.5576923076923076e-05,
|
| 2156 |
+
"loss": 2.2891,
|
| 2157 |
+
"step": 2148
|
| 2158 |
+
},
|
| 2159 |
+
{
|
| 2160 |
+
"epoch": 6.9,
|
| 2161 |
+
"learning_rate": 1.5480769230769232e-05,
|
| 2162 |
+
"loss": 2.3154,
|
| 2163 |
+
"step": 2154
|
| 2164 |
+
},
|
| 2165 |
+
{
|
| 2166 |
+
"epoch": 6.92,
|
| 2167 |
+
"learning_rate": 1.5384615384615387e-05,
|
| 2168 |
+
"loss": 2.3642,
|
| 2169 |
+
"step": 2160
|
| 2170 |
+
},
|
| 2171 |
+
{
|
| 2172 |
+
"epoch": 6.94,
|
| 2173 |
+
"learning_rate": 1.528846153846154e-05,
|
| 2174 |
+
"loss": 2.3303,
|
| 2175 |
+
"step": 2166
|
| 2176 |
+
},
|
| 2177 |
+
{
|
| 2178 |
+
"epoch": 6.96,
|
| 2179 |
+
"learning_rate": 1.5192307692307691e-05,
|
| 2180 |
+
"loss": 2.4364,
|
| 2181 |
+
"step": 2172
|
| 2182 |
+
},
|
| 2183 |
+
{
|
| 2184 |
+
"epoch": 6.98,
|
| 2185 |
+
"learning_rate": 1.5096153846153847e-05,
|
| 2186 |
+
"loss": 2.4258,
|
| 2187 |
+
"step": 2178
|
| 2188 |
+
},
|
| 2189 |
+
{
|
| 2190 |
+
"epoch": 7.0,
|
| 2191 |
+
"learning_rate": 1.5e-05,
|
| 2192 |
+
"loss": 2.3566,
|
| 2193 |
+
"step": 2184
|
| 2194 |
+
},
|
| 2195 |
+
{
|
| 2196 |
+
"epoch": 7.02,
|
| 2197 |
+
"learning_rate": 1.4903846153846154e-05,
|
| 2198 |
+
"loss": 2.2916,
|
| 2199 |
+
"step": 2190
|
| 2200 |
+
},
|
| 2201 |
+
{
|
| 2202 |
+
"epoch": 7.04,
|
| 2203 |
+
"learning_rate": 1.4807692307692308e-05,
|
| 2204 |
+
"loss": 2.3829,
|
| 2205 |
+
"step": 2196
|
| 2206 |
+
},
|
| 2207 |
+
{
|
| 2208 |
+
"epoch": 7.06,
|
| 2209 |
+
"learning_rate": 1.4711538461538463e-05,
|
| 2210 |
+
"loss": 2.242,
|
| 2211 |
+
"step": 2202
|
| 2212 |
+
},
|
| 2213 |
+
{
|
| 2214 |
+
"epoch": 7.08,
|
| 2215 |
+
"learning_rate": 1.4615384615384617e-05,
|
| 2216 |
+
"loss": 2.2546,
|
| 2217 |
+
"step": 2208
|
| 2218 |
+
},
|
| 2219 |
+
{
|
| 2220 |
+
"epoch": 7.1,
|
| 2221 |
+
"learning_rate": 1.451923076923077e-05,
|
| 2222 |
+
"loss": 2.1526,
|
| 2223 |
+
"step": 2214
|
| 2224 |
+
},
|
| 2225 |
+
{
|
| 2226 |
+
"epoch": 7.12,
|
| 2227 |
+
"learning_rate": 1.4423076923076923e-05,
|
| 2228 |
+
"loss": 2.2073,
|
| 2229 |
+
"step": 2220
|
| 2230 |
+
},
|
| 2231 |
+
{
|
| 2232 |
+
"epoch": 7.13,
|
| 2233 |
+
"learning_rate": 1.4326923076923076e-05,
|
| 2234 |
+
"loss": 2.2992,
|
| 2235 |
+
"step": 2226
|
| 2236 |
+
},
|
| 2237 |
+
{
|
| 2238 |
+
"epoch": 7.15,
|
| 2239 |
+
"learning_rate": 1.423076923076923e-05,
|
| 2240 |
+
"loss": 2.2544,
|
| 2241 |
+
"step": 2232
|
| 2242 |
+
},
|
| 2243 |
+
{
|
| 2244 |
+
"epoch": 7.17,
|
| 2245 |
+
"learning_rate": 1.4134615384615386e-05,
|
| 2246 |
+
"loss": 2.2632,
|
| 2247 |
+
"step": 2238
|
| 2248 |
+
},
|
| 2249 |
+
{
|
| 2250 |
+
"epoch": 7.19,
|
| 2251 |
+
"learning_rate": 1.403846153846154e-05,
|
| 2252 |
+
"loss": 2.3158,
|
| 2253 |
+
"step": 2244
|
| 2254 |
+
},
|
| 2255 |
+
{
|
| 2256 |
+
"epoch": 7.21,
|
| 2257 |
+
"learning_rate": 1.3942307692307693e-05,
|
| 2258 |
+
"loss": 2.2981,
|
| 2259 |
+
"step": 2250
|
| 2260 |
+
},
|
| 2261 |
+
{
|
| 2262 |
+
"epoch": 7.23,
|
| 2263 |
+
"learning_rate": 1.3846153846153847e-05,
|
| 2264 |
+
"loss": 2.2468,
|
| 2265 |
+
"step": 2256
|
| 2266 |
+
},
|
| 2267 |
+
{
|
| 2268 |
+
"epoch": 7.25,
|
| 2269 |
+
"learning_rate": 1.3750000000000002e-05,
|
| 2270 |
+
"loss": 2.4333,
|
| 2271 |
+
"step": 2262
|
| 2272 |
+
},
|
| 2273 |
+
{
|
| 2274 |
+
"epoch": 7.27,
|
| 2275 |
+
"learning_rate": 1.3653846153846153e-05,
|
| 2276 |
+
"loss": 2.3005,
|
| 2277 |
+
"step": 2268
|
| 2278 |
+
},
|
| 2279 |
+
{
|
| 2280 |
+
"epoch": 7.29,
|
| 2281 |
+
"learning_rate": 1.3557692307692308e-05,
|
| 2282 |
+
"loss": 2.2629,
|
| 2283 |
+
"step": 2274
|
| 2284 |
+
},
|
| 2285 |
+
{
|
| 2286 |
+
"epoch": 7.31,
|
| 2287 |
+
"learning_rate": 1.3461538461538462e-05,
|
| 2288 |
+
"loss": 2.3122,
|
| 2289 |
+
"step": 2280
|
| 2290 |
+
},
|
| 2291 |
+
{
|
| 2292 |
+
"epoch": 7.33,
|
| 2293 |
+
"learning_rate": 1.3365384615384615e-05,
|
| 2294 |
+
"loss": 2.3257,
|
| 2295 |
+
"step": 2286
|
| 2296 |
+
},
|
| 2297 |
+
{
|
| 2298 |
+
"epoch": 7.35,
|
| 2299 |
+
"learning_rate": 1.3269230769230769e-05,
|
| 2300 |
+
"loss": 2.4191,
|
| 2301 |
+
"step": 2292
|
| 2302 |
+
},
|
| 2303 |
+
{
|
| 2304 |
+
"epoch": 7.37,
|
| 2305 |
+
"learning_rate": 1.3173076923076925e-05,
|
| 2306 |
+
"loss": 2.2431,
|
| 2307 |
+
"step": 2298
|
| 2308 |
+
},
|
| 2309 |
+
{
|
| 2310 |
+
"epoch": 7.38,
|
| 2311 |
+
"learning_rate": 1.3076923076923078e-05,
|
| 2312 |
+
"loss": 2.3716,
|
| 2313 |
+
"step": 2304
|
| 2314 |
+
},
|
| 2315 |
+
{
|
| 2316 |
+
"epoch": 7.4,
|
| 2317 |
+
"learning_rate": 1.2980769230769232e-05,
|
| 2318 |
+
"loss": 2.1685,
|
| 2319 |
+
"step": 2310
|
| 2320 |
+
},
|
| 2321 |
+
{
|
| 2322 |
+
"epoch": 7.42,
|
| 2323 |
+
"learning_rate": 1.2884615384615384e-05,
|
| 2324 |
+
"loss": 2.2519,
|
| 2325 |
+
"step": 2316
|
| 2326 |
+
},
|
| 2327 |
+
{
|
| 2328 |
+
"epoch": 7.44,
|
| 2329 |
+
"learning_rate": 1.2788461538461538e-05,
|
| 2330 |
+
"loss": 2.2242,
|
| 2331 |
+
"step": 2322
|
| 2332 |
+
},
|
| 2333 |
+
{
|
| 2334 |
+
"epoch": 7.46,
|
| 2335 |
+
"learning_rate": 1.2692307692307691e-05,
|
| 2336 |
+
"loss": 2.3481,
|
| 2337 |
+
"step": 2328
|
| 2338 |
+
},
|
| 2339 |
+
{
|
| 2340 |
+
"epoch": 7.48,
|
| 2341 |
+
"learning_rate": 1.2596153846153847e-05,
|
| 2342 |
+
"loss": 2.3117,
|
| 2343 |
+
"step": 2334
|
| 2344 |
+
},
|
| 2345 |
+
{
|
| 2346 |
+
"epoch": 7.5,
|
| 2347 |
+
"learning_rate": 1.25e-05,
|
| 2348 |
+
"loss": 2.2924,
|
| 2349 |
+
"step": 2340
|
| 2350 |
+
},
|
| 2351 |
+
{
|
| 2352 |
+
"epoch": 7.52,
|
| 2353 |
+
"learning_rate": 1.2403846153846154e-05,
|
| 2354 |
+
"loss": 2.2933,
|
| 2355 |
+
"step": 2346
|
| 2356 |
+
},
|
| 2357 |
+
{
|
| 2358 |
+
"epoch": 7.54,
|
| 2359 |
+
"learning_rate": 1.230769230769231e-05,
|
| 2360 |
+
"loss": 2.298,
|
| 2361 |
+
"step": 2352
|
| 2362 |
+
},
|
| 2363 |
+
{
|
| 2364 |
+
"epoch": 7.56,
|
| 2365 |
+
"learning_rate": 1.2211538461538462e-05,
|
| 2366 |
+
"loss": 2.2145,
|
| 2367 |
+
"step": 2358
|
| 2368 |
+
},
|
| 2369 |
+
{
|
| 2370 |
+
"epoch": 7.58,
|
| 2371 |
+
"learning_rate": 1.2115384615384615e-05,
|
| 2372 |
+
"loss": 2.3118,
|
| 2373 |
+
"step": 2364
|
| 2374 |
+
},
|
| 2375 |
+
{
|
| 2376 |
+
"epoch": 7.6,
|
| 2377 |
+
"learning_rate": 1.2019230769230771e-05,
|
| 2378 |
+
"loss": 2.3132,
|
| 2379 |
+
"step": 2370
|
| 2380 |
+
},
|
| 2381 |
+
{
|
| 2382 |
+
"epoch": 7.62,
|
| 2383 |
+
"learning_rate": 1.1923076923076925e-05,
|
| 2384 |
+
"loss": 2.4653,
|
| 2385 |
+
"step": 2376
|
| 2386 |
+
},
|
| 2387 |
+
{
|
| 2388 |
+
"epoch": 7.63,
|
| 2389 |
+
"learning_rate": 1.1826923076923077e-05,
|
| 2390 |
+
"loss": 2.2083,
|
| 2391 |
+
"step": 2382
|
| 2392 |
+
},
|
| 2393 |
+
{
|
| 2394 |
+
"epoch": 7.65,
|
| 2395 |
+
"learning_rate": 1.1730769230769232e-05,
|
| 2396 |
+
"loss": 2.2876,
|
| 2397 |
+
"step": 2388
|
| 2398 |
+
},
|
| 2399 |
+
{
|
| 2400 |
+
"epoch": 7.67,
|
| 2401 |
+
"learning_rate": 1.1634615384615386e-05,
|
| 2402 |
+
"loss": 2.2115,
|
| 2403 |
+
"step": 2394
|
| 2404 |
+
},
|
| 2405 |
+
{
|
| 2406 |
+
"epoch": 7.69,
|
| 2407 |
+
"learning_rate": 1.153846153846154e-05,
|
| 2408 |
+
"loss": 2.3116,
|
| 2409 |
+
"step": 2400
|
| 2410 |
+
},
|
| 2411 |
+
{
|
| 2412 |
+
"epoch": 7.71,
|
| 2413 |
+
"learning_rate": 1.1442307692307693e-05,
|
| 2414 |
+
"loss": 2.2858,
|
| 2415 |
+
"step": 2406
|
| 2416 |
+
},
|
| 2417 |
+
{
|
| 2418 |
+
"epoch": 7.73,
|
| 2419 |
+
"learning_rate": 1.1346153846153847e-05,
|
| 2420 |
+
"loss": 2.2525,
|
| 2421 |
+
"step": 2412
|
| 2422 |
+
},
|
| 2423 |
+
{
|
| 2424 |
+
"epoch": 7.75,
|
| 2425 |
+
"learning_rate": 1.125e-05,
|
| 2426 |
+
"loss": 2.3215,
|
| 2427 |
+
"step": 2418
|
| 2428 |
+
},
|
| 2429 |
+
{
|
| 2430 |
+
"epoch": 7.77,
|
| 2431 |
+
"learning_rate": 1.1153846153846154e-05,
|
| 2432 |
+
"loss": 2.3884,
|
| 2433 |
+
"step": 2424
|
| 2434 |
+
},
|
| 2435 |
+
{
|
| 2436 |
+
"epoch": 7.79,
|
| 2437 |
+
"learning_rate": 1.1057692307692308e-05,
|
| 2438 |
+
"loss": 2.3767,
|
| 2439 |
+
"step": 2430
|
| 2440 |
+
},
|
| 2441 |
+
{
|
| 2442 |
+
"epoch": 7.81,
|
| 2443 |
+
"learning_rate": 1.0961538461538462e-05,
|
| 2444 |
+
"loss": 2.1214,
|
| 2445 |
+
"step": 2436
|
| 2446 |
+
},
|
| 2447 |
+
{
|
| 2448 |
+
"epoch": 7.83,
|
| 2449 |
+
"learning_rate": 1.0865384615384616e-05,
|
| 2450 |
+
"loss": 2.3003,
|
| 2451 |
+
"step": 2442
|
| 2452 |
+
},
|
| 2453 |
+
{
|
| 2454 |
+
"epoch": 7.85,
|
| 2455 |
+
"learning_rate": 1.0769230769230771e-05,
|
| 2456 |
+
"loss": 2.426,
|
| 2457 |
+
"step": 2448
|
| 2458 |
+
},
|
| 2459 |
+
{
|
| 2460 |
+
"epoch": 7.87,
|
| 2461 |
+
"learning_rate": 1.0673076923076923e-05,
|
| 2462 |
+
"loss": 2.2785,
|
| 2463 |
+
"step": 2454
|
| 2464 |
+
},
|
| 2465 |
+
{
|
| 2466 |
+
"epoch": 7.88,
|
| 2467 |
+
"learning_rate": 1.0576923076923077e-05,
|
| 2468 |
+
"loss": 2.3252,
|
| 2469 |
+
"step": 2460
|
| 2470 |
+
},
|
| 2471 |
+
{
|
| 2472 |
+
"epoch": 7.9,
|
| 2473 |
+
"learning_rate": 1.0480769230769232e-05,
|
| 2474 |
+
"loss": 2.2636,
|
| 2475 |
+
"step": 2466
|
| 2476 |
+
},
|
| 2477 |
+
{
|
| 2478 |
+
"epoch": 7.92,
|
| 2479 |
+
"learning_rate": 1.0384615384615386e-05,
|
| 2480 |
+
"loss": 2.2732,
|
| 2481 |
+
"step": 2472
|
| 2482 |
+
},
|
| 2483 |
+
{
|
| 2484 |
+
"epoch": 7.94,
|
| 2485 |
+
"learning_rate": 1.0288461538461538e-05,
|
| 2486 |
+
"loss": 2.1934,
|
| 2487 |
+
"step": 2478
|
| 2488 |
+
},
|
| 2489 |
+
{
|
| 2490 |
+
"epoch": 7.96,
|
| 2491 |
+
"learning_rate": 1.0192307692307693e-05,
|
| 2492 |
+
"loss": 2.2967,
|
| 2493 |
+
"step": 2484
|
| 2494 |
+
},
|
| 2495 |
+
{
|
| 2496 |
+
"epoch": 7.98,
|
| 2497 |
+
"learning_rate": 1.0096153846153847e-05,
|
| 2498 |
+
"loss": 2.2939,
|
| 2499 |
+
"step": 2490
|
| 2500 |
+
},
|
| 2501 |
+
{
|
| 2502 |
+
"epoch": 8.0,
|
| 2503 |
+
"learning_rate": 1e-05,
|
| 2504 |
+
"loss": 2.3393,
|
| 2505 |
+
"step": 2496
|
| 2506 |
+
}
|
| 2507 |
+
],
|
| 2508 |
+
"logging_steps": 6,
|
| 2509 |
+
"max_steps": 3120,
|
| 2510 |
+
"num_train_epochs": 10,
|
| 2511 |
+
"save_steps": 500,
|
| 2512 |
+
"total_flos": 1007380809216000.0,
|
| 2513 |
+
"trial_name": null,
|
| 2514 |
+
"trial_params": null
|
| 2515 |
+
}
|
checkpoint-2500/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c150481b2be1a59fc2f8a97e3f83518bf0fe89d39fbfc7d3d90a4bdd3f30ccaf
|
| 3 |
+
size 4600
|
checkpoint-3000/config.json
ADDED
|
@@ -0,0 +1,40 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "flax-community/gpt2-small-indonesian",
|
| 3 |
+
"activation_function": "gelu_new",
|
| 4 |
+
"architectures": [
|
| 5 |
+
"GPT2LMHeadModel"
|
| 6 |
+
],
|
| 7 |
+
"attn_pdrop": 0.0,
|
| 8 |
+
"bos_token_id": 50256,
|
| 9 |
+
"embd_pdrop": 0.0,
|
| 10 |
+
"eos_token_id": 50256,
|
| 11 |
+
"gradient_checkpointing": false,
|
| 12 |
+
"initializer_range": 0.02,
|
| 13 |
+
"layer_norm_epsilon": 1e-05,
|
| 14 |
+
"model_type": "gpt2",
|
| 15 |
+
"n_ctx": 1024,
|
| 16 |
+
"n_embd": 768,
|
| 17 |
+
"n_head": 12,
|
| 18 |
+
"n_inner": null,
|
| 19 |
+
"n_layer": 12,
|
| 20 |
+
"n_positions": 1024,
|
| 21 |
+
"reorder_and_upcast_attn": false,
|
| 22 |
+
"resid_pdrop": 0.0,
|
| 23 |
+
"scale_attn_by_inverse_layer_idx": false,
|
| 24 |
+
"scale_attn_weights": true,
|
| 25 |
+
"summary_activation": null,
|
| 26 |
+
"summary_first_dropout": 0.1,
|
| 27 |
+
"summary_proj_to_labels": true,
|
| 28 |
+
"summary_type": "cls_index",
|
| 29 |
+
"summary_use_proj": true,
|
| 30 |
+
"task_specific_params": {
|
| 31 |
+
"text-generation": {
|
| 32 |
+
"do_sample": true,
|
| 33 |
+
"max_length": 50
|
| 34 |
+
}
|
| 35 |
+
},
|
| 36 |
+
"torch_dtype": "float32",
|
| 37 |
+
"transformers_version": "4.35.2",
|
| 38 |
+
"use_cache": true,
|
| 39 |
+
"vocab_size": 50260
|
| 40 |
+
}
|
checkpoint-3000/generation_config.json
ADDED
|
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"bos_token_id": 50256,
|
| 4 |
+
"eos_token_id": 50256,
|
| 5 |
+
"transformers_version": "4.35.2"
|
| 6 |
+
}
|
checkpoint-3000/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7f58bef7a816dd82612425a9d5bdba5dc74e8975f6fe3012d981d3c4e51cf384
|
| 3 |
+
size 497783424
|
checkpoint-3000/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:fbff0ff8f6cb0844dd53eba8f286c3834886c6c5cab5366b7b318a412ef4e7be
|
| 3 |
+
size 995660730
|
checkpoint-3000/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9afbf853352cfbcfea61884ff6a2ddcd2aee1ce8618589cf5b56912c1b160011
|
| 3 |
+
size 14244
|
checkpoint-3000/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0c1c221b1dc86b6645f8fae1499e982f46b4d6f8f14aa5c67e95ba7f117afa44
|
| 3 |
+
size 1064
|
checkpoint-3000/trainer_state.json
ADDED
|
@@ -0,0 +1,3019 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 9.615384615384615,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 3000,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.02,
|
| 13 |
+
"learning_rate": 4.9903846153846154e-05,
|
| 14 |
+
"loss": 5.8218,
|
| 15 |
+
"step": 6
|
| 16 |
+
},
|
| 17 |
+
{
|
| 18 |
+
"epoch": 0.04,
|
| 19 |
+
"learning_rate": 4.980769230769231e-05,
|
| 20 |
+
"loss": 5.2499,
|
| 21 |
+
"step": 12
|
| 22 |
+
},
|
| 23 |
+
{
|
| 24 |
+
"epoch": 0.06,
|
| 25 |
+
"learning_rate": 4.9711538461538465e-05,
|
| 26 |
+
"loss": 5.1236,
|
| 27 |
+
"step": 18
|
| 28 |
+
},
|
| 29 |
+
{
|
| 30 |
+
"epoch": 0.08,
|
| 31 |
+
"learning_rate": 4.961538461538462e-05,
|
| 32 |
+
"loss": 4.8363,
|
| 33 |
+
"step": 24
|
| 34 |
+
},
|
| 35 |
+
{
|
| 36 |
+
"epoch": 0.1,
|
| 37 |
+
"learning_rate": 4.9519230769230776e-05,
|
| 38 |
+
"loss": 4.5805,
|
| 39 |
+
"step": 30
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.12,
|
| 43 |
+
"learning_rate": 4.942307692307693e-05,
|
| 44 |
+
"loss": 4.5276,
|
| 45 |
+
"step": 36
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.13,
|
| 49 |
+
"learning_rate": 4.932692307692308e-05,
|
| 50 |
+
"loss": 4.3871,
|
| 51 |
+
"step": 42
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.15,
|
| 55 |
+
"learning_rate": 4.923076923076924e-05,
|
| 56 |
+
"loss": 4.2706,
|
| 57 |
+
"step": 48
|
| 58 |
+
},
|
| 59 |
+
{
|
| 60 |
+
"epoch": 0.17,
|
| 61 |
+
"learning_rate": 4.9134615384615384e-05,
|
| 62 |
+
"loss": 4.0906,
|
| 63 |
+
"step": 54
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 0.19,
|
| 67 |
+
"learning_rate": 4.9038461538461536e-05,
|
| 68 |
+
"loss": 4.1704,
|
| 69 |
+
"step": 60
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 0.21,
|
| 73 |
+
"learning_rate": 4.8942307692307695e-05,
|
| 74 |
+
"loss": 3.9014,
|
| 75 |
+
"step": 66
|
| 76 |
+
},
|
| 77 |
+
{
|
| 78 |
+
"epoch": 0.23,
|
| 79 |
+
"learning_rate": 4.884615384615385e-05,
|
| 80 |
+
"loss": 4.0338,
|
| 81 |
+
"step": 72
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"epoch": 0.25,
|
| 85 |
+
"learning_rate": 4.875e-05,
|
| 86 |
+
"loss": 4.011,
|
| 87 |
+
"step": 78
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"epoch": 0.27,
|
| 91 |
+
"learning_rate": 4.865384615384616e-05,
|
| 92 |
+
"loss": 3.8017,
|
| 93 |
+
"step": 84
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.29,
|
| 97 |
+
"learning_rate": 4.855769230769231e-05,
|
| 98 |
+
"loss": 3.819,
|
| 99 |
+
"step": 90
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"epoch": 0.31,
|
| 103 |
+
"learning_rate": 4.846153846153846e-05,
|
| 104 |
+
"loss": 3.8157,
|
| 105 |
+
"step": 96
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 0.33,
|
| 109 |
+
"learning_rate": 4.836538461538462e-05,
|
| 110 |
+
"loss": 3.7675,
|
| 111 |
+
"step": 102
|
| 112 |
+
},
|
| 113 |
+
{
|
| 114 |
+
"epoch": 0.35,
|
| 115 |
+
"learning_rate": 4.826923076923077e-05,
|
| 116 |
+
"loss": 3.756,
|
| 117 |
+
"step": 108
|
| 118 |
+
},
|
| 119 |
+
{
|
| 120 |
+
"epoch": 0.37,
|
| 121 |
+
"learning_rate": 4.8173076923076925e-05,
|
| 122 |
+
"loss": 3.7439,
|
| 123 |
+
"step": 114
|
| 124 |
+
},
|
| 125 |
+
{
|
| 126 |
+
"epoch": 0.38,
|
| 127 |
+
"learning_rate": 4.8076923076923084e-05,
|
| 128 |
+
"loss": 3.761,
|
| 129 |
+
"step": 120
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.4,
|
| 133 |
+
"learning_rate": 4.7980769230769236e-05,
|
| 134 |
+
"loss": 3.7318,
|
| 135 |
+
"step": 126
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.42,
|
| 139 |
+
"learning_rate": 4.788461538461539e-05,
|
| 140 |
+
"loss": 3.745,
|
| 141 |
+
"step": 132
|
| 142 |
+
},
|
| 143 |
+
{
|
| 144 |
+
"epoch": 0.44,
|
| 145 |
+
"learning_rate": 4.778846153846154e-05,
|
| 146 |
+
"loss": 3.784,
|
| 147 |
+
"step": 138
|
| 148 |
+
},
|
| 149 |
+
{
|
| 150 |
+
"epoch": 0.46,
|
| 151 |
+
"learning_rate": 4.76923076923077e-05,
|
| 152 |
+
"loss": 3.6114,
|
| 153 |
+
"step": 144
|
| 154 |
+
},
|
| 155 |
+
{
|
| 156 |
+
"epoch": 0.48,
|
| 157 |
+
"learning_rate": 4.7596153846153844e-05,
|
| 158 |
+
"loss": 3.637,
|
| 159 |
+
"step": 150
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 0.5,
|
| 163 |
+
"learning_rate": 4.75e-05,
|
| 164 |
+
"loss": 3.5881,
|
| 165 |
+
"step": 156
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 0.52,
|
| 169 |
+
"learning_rate": 4.7403846153846155e-05,
|
| 170 |
+
"loss": 3.5336,
|
| 171 |
+
"step": 162
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 0.54,
|
| 175 |
+
"learning_rate": 4.730769230769231e-05,
|
| 176 |
+
"loss": 3.5313,
|
| 177 |
+
"step": 168
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.56,
|
| 181 |
+
"learning_rate": 4.7211538461538465e-05,
|
| 182 |
+
"loss": 3.6357,
|
| 183 |
+
"step": 174
|
| 184 |
+
},
|
| 185 |
+
{
|
| 186 |
+
"epoch": 0.58,
|
| 187 |
+
"learning_rate": 4.711538461538462e-05,
|
| 188 |
+
"loss": 3.5818,
|
| 189 |
+
"step": 180
|
| 190 |
+
},
|
| 191 |
+
{
|
| 192 |
+
"epoch": 0.6,
|
| 193 |
+
"learning_rate": 4.701923076923077e-05,
|
| 194 |
+
"loss": 3.5017,
|
| 195 |
+
"step": 186
|
| 196 |
+
},
|
| 197 |
+
{
|
| 198 |
+
"epoch": 0.62,
|
| 199 |
+
"learning_rate": 4.692307692307693e-05,
|
| 200 |
+
"loss": 3.5815,
|
| 201 |
+
"step": 192
|
| 202 |
+
},
|
| 203 |
+
{
|
| 204 |
+
"epoch": 0.63,
|
| 205 |
+
"learning_rate": 4.682692307692308e-05,
|
| 206 |
+
"loss": 3.611,
|
| 207 |
+
"step": 198
|
| 208 |
+
},
|
| 209 |
+
{
|
| 210 |
+
"epoch": 0.65,
|
| 211 |
+
"learning_rate": 4.673076923076923e-05,
|
| 212 |
+
"loss": 3.6096,
|
| 213 |
+
"step": 204
|
| 214 |
+
},
|
| 215 |
+
{
|
| 216 |
+
"epoch": 0.67,
|
| 217 |
+
"learning_rate": 4.6634615384615384e-05,
|
| 218 |
+
"loss": 3.5436,
|
| 219 |
+
"step": 210
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.69,
|
| 223 |
+
"learning_rate": 4.653846153846154e-05,
|
| 224 |
+
"loss": 3.5258,
|
| 225 |
+
"step": 216
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 0.71,
|
| 229 |
+
"learning_rate": 4.6442307692307695e-05,
|
| 230 |
+
"loss": 3.5277,
|
| 231 |
+
"step": 222
|
| 232 |
+
},
|
| 233 |
+
{
|
| 234 |
+
"epoch": 0.73,
|
| 235 |
+
"learning_rate": 4.634615384615385e-05,
|
| 236 |
+
"loss": 3.505,
|
| 237 |
+
"step": 228
|
| 238 |
+
},
|
| 239 |
+
{
|
| 240 |
+
"epoch": 0.75,
|
| 241 |
+
"learning_rate": 4.6250000000000006e-05,
|
| 242 |
+
"loss": 3.4665,
|
| 243 |
+
"step": 234
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"epoch": 0.77,
|
| 247 |
+
"learning_rate": 4.615384615384616e-05,
|
| 248 |
+
"loss": 3.4245,
|
| 249 |
+
"step": 240
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 0.79,
|
| 253 |
+
"learning_rate": 4.605769230769231e-05,
|
| 254 |
+
"loss": 3.4547,
|
| 255 |
+
"step": 246
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 0.81,
|
| 259 |
+
"learning_rate": 4.596153846153846e-05,
|
| 260 |
+
"loss": 3.4382,
|
| 261 |
+
"step": 252
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.83,
|
| 265 |
+
"learning_rate": 4.5865384615384614e-05,
|
| 266 |
+
"loss": 3.482,
|
| 267 |
+
"step": 258
|
| 268 |
+
},
|
| 269 |
+
{
|
| 270 |
+
"epoch": 0.85,
|
| 271 |
+
"learning_rate": 4.576923076923077e-05,
|
| 272 |
+
"loss": 3.4795,
|
| 273 |
+
"step": 264
|
| 274 |
+
},
|
| 275 |
+
{
|
| 276 |
+
"epoch": 0.87,
|
| 277 |
+
"learning_rate": 4.5673076923076925e-05,
|
| 278 |
+
"loss": 3.5018,
|
| 279 |
+
"step": 270
|
| 280 |
+
},
|
| 281 |
+
{
|
| 282 |
+
"epoch": 0.88,
|
| 283 |
+
"learning_rate": 4.557692307692308e-05,
|
| 284 |
+
"loss": 3.448,
|
| 285 |
+
"step": 276
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 0.9,
|
| 289 |
+
"learning_rate": 4.548076923076923e-05,
|
| 290 |
+
"loss": 3.4206,
|
| 291 |
+
"step": 282
|
| 292 |
+
},
|
| 293 |
+
{
|
| 294 |
+
"epoch": 0.92,
|
| 295 |
+
"learning_rate": 4.538461538461539e-05,
|
| 296 |
+
"loss": 3.528,
|
| 297 |
+
"step": 288
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 0.94,
|
| 301 |
+
"learning_rate": 4.528846153846154e-05,
|
| 302 |
+
"loss": 3.475,
|
| 303 |
+
"step": 294
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.96,
|
| 307 |
+
"learning_rate": 4.519230769230769e-05,
|
| 308 |
+
"loss": 3.4142,
|
| 309 |
+
"step": 300
|
| 310 |
+
},
|
| 311 |
+
{
|
| 312 |
+
"epoch": 0.98,
|
| 313 |
+
"learning_rate": 4.509615384615385e-05,
|
| 314 |
+
"loss": 3.4156,
|
| 315 |
+
"step": 306
|
| 316 |
+
},
|
| 317 |
+
{
|
| 318 |
+
"epoch": 1.0,
|
| 319 |
+
"learning_rate": 4.5e-05,
|
| 320 |
+
"loss": 3.3274,
|
| 321 |
+
"step": 312
|
| 322 |
+
},
|
| 323 |
+
{
|
| 324 |
+
"epoch": 1.02,
|
| 325 |
+
"learning_rate": 4.4903846153846155e-05,
|
| 326 |
+
"loss": 3.3594,
|
| 327 |
+
"step": 318
|
| 328 |
+
},
|
| 329 |
+
{
|
| 330 |
+
"epoch": 1.04,
|
| 331 |
+
"learning_rate": 4.4807692307692314e-05,
|
| 332 |
+
"loss": 3.2517,
|
| 333 |
+
"step": 324
|
| 334 |
+
},
|
| 335 |
+
{
|
| 336 |
+
"epoch": 1.06,
|
| 337 |
+
"learning_rate": 4.4711538461538466e-05,
|
| 338 |
+
"loss": 3.2934,
|
| 339 |
+
"step": 330
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 1.08,
|
| 343 |
+
"learning_rate": 4.461538461538462e-05,
|
| 344 |
+
"loss": 3.2986,
|
| 345 |
+
"step": 336
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 1.1,
|
| 349 |
+
"learning_rate": 4.451923076923077e-05,
|
| 350 |
+
"loss": 3.2655,
|
| 351 |
+
"step": 342
|
| 352 |
+
},
|
| 353 |
+
{
|
| 354 |
+
"epoch": 1.12,
|
| 355 |
+
"learning_rate": 4.442307692307692e-05,
|
| 356 |
+
"loss": 3.4058,
|
| 357 |
+
"step": 348
|
| 358 |
+
},
|
| 359 |
+
{
|
| 360 |
+
"epoch": 1.13,
|
| 361 |
+
"learning_rate": 4.4326923076923074e-05,
|
| 362 |
+
"loss": 3.265,
|
| 363 |
+
"step": 354
|
| 364 |
+
},
|
| 365 |
+
{
|
| 366 |
+
"epoch": 1.15,
|
| 367 |
+
"learning_rate": 4.423076923076923e-05,
|
| 368 |
+
"loss": 3.3208,
|
| 369 |
+
"step": 360
|
| 370 |
+
},
|
| 371 |
+
{
|
| 372 |
+
"epoch": 1.17,
|
| 373 |
+
"learning_rate": 4.4134615384615385e-05,
|
| 374 |
+
"loss": 3.238,
|
| 375 |
+
"step": 366
|
| 376 |
+
},
|
| 377 |
+
{
|
| 378 |
+
"epoch": 1.19,
|
| 379 |
+
"learning_rate": 4.403846153846154e-05,
|
| 380 |
+
"loss": 3.2153,
|
| 381 |
+
"step": 372
|
| 382 |
+
},
|
| 383 |
+
{
|
| 384 |
+
"epoch": 1.21,
|
| 385 |
+
"learning_rate": 4.3942307692307695e-05,
|
| 386 |
+
"loss": 3.264,
|
| 387 |
+
"step": 378
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 1.23,
|
| 391 |
+
"learning_rate": 4.384615384615385e-05,
|
| 392 |
+
"loss": 3.2911,
|
| 393 |
+
"step": 384
|
| 394 |
+
},
|
| 395 |
+
{
|
| 396 |
+
"epoch": 1.25,
|
| 397 |
+
"learning_rate": 4.375e-05,
|
| 398 |
+
"loss": 3.3027,
|
| 399 |
+
"step": 390
|
| 400 |
+
},
|
| 401 |
+
{
|
| 402 |
+
"epoch": 1.27,
|
| 403 |
+
"learning_rate": 4.365384615384616e-05,
|
| 404 |
+
"loss": 3.2589,
|
| 405 |
+
"step": 396
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 1.29,
|
| 409 |
+
"learning_rate": 4.355769230769231e-05,
|
| 410 |
+
"loss": 3.3683,
|
| 411 |
+
"step": 402
|
| 412 |
+
},
|
| 413 |
+
{
|
| 414 |
+
"epoch": 1.31,
|
| 415 |
+
"learning_rate": 4.346153846153846e-05,
|
| 416 |
+
"loss": 3.2849,
|
| 417 |
+
"step": 408
|
| 418 |
+
},
|
| 419 |
+
{
|
| 420 |
+
"epoch": 1.33,
|
| 421 |
+
"learning_rate": 4.336538461538462e-05,
|
| 422 |
+
"loss": 3.2397,
|
| 423 |
+
"step": 414
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 1.35,
|
| 427 |
+
"learning_rate": 4.326923076923077e-05,
|
| 428 |
+
"loss": 3.2128,
|
| 429 |
+
"step": 420
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 1.37,
|
| 433 |
+
"learning_rate": 4.3173076923076925e-05,
|
| 434 |
+
"loss": 3.1944,
|
| 435 |
+
"step": 426
|
| 436 |
+
},
|
| 437 |
+
{
|
| 438 |
+
"epoch": 1.38,
|
| 439 |
+
"learning_rate": 4.3076923076923084e-05,
|
| 440 |
+
"loss": 3.1837,
|
| 441 |
+
"step": 432
|
| 442 |
+
},
|
| 443 |
+
{
|
| 444 |
+
"epoch": 1.4,
|
| 445 |
+
"learning_rate": 4.2980769230769236e-05,
|
| 446 |
+
"loss": 3.1793,
|
| 447 |
+
"step": 438
|
| 448 |
+
},
|
| 449 |
+
{
|
| 450 |
+
"epoch": 1.42,
|
| 451 |
+
"learning_rate": 4.288461538461538e-05,
|
| 452 |
+
"loss": 3.1447,
|
| 453 |
+
"step": 444
|
| 454 |
+
},
|
| 455 |
+
{
|
| 456 |
+
"epoch": 1.44,
|
| 457 |
+
"learning_rate": 4.278846153846154e-05,
|
| 458 |
+
"loss": 3.1028,
|
| 459 |
+
"step": 450
|
| 460 |
+
},
|
| 461 |
+
{
|
| 462 |
+
"epoch": 1.46,
|
| 463 |
+
"learning_rate": 4.269230769230769e-05,
|
| 464 |
+
"loss": 3.2471,
|
| 465 |
+
"step": 456
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 1.48,
|
| 469 |
+
"learning_rate": 4.2596153846153844e-05,
|
| 470 |
+
"loss": 3.1855,
|
| 471 |
+
"step": 462
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 1.5,
|
| 475 |
+
"learning_rate": 4.25e-05,
|
| 476 |
+
"loss": 3.1817,
|
| 477 |
+
"step": 468
|
| 478 |
+
},
|
| 479 |
+
{
|
| 480 |
+
"epoch": 1.52,
|
| 481 |
+
"learning_rate": 4.2403846153846155e-05,
|
| 482 |
+
"loss": 3.214,
|
| 483 |
+
"step": 474
|
| 484 |
+
},
|
| 485 |
+
{
|
| 486 |
+
"epoch": 1.54,
|
| 487 |
+
"learning_rate": 4.230769230769231e-05,
|
| 488 |
+
"loss": 2.997,
|
| 489 |
+
"step": 480
|
| 490 |
+
},
|
| 491 |
+
{
|
| 492 |
+
"epoch": 1.56,
|
| 493 |
+
"learning_rate": 4.2211538461538466e-05,
|
| 494 |
+
"loss": 3.2059,
|
| 495 |
+
"step": 486
|
| 496 |
+
},
|
| 497 |
+
{
|
| 498 |
+
"epoch": 1.58,
|
| 499 |
+
"learning_rate": 4.211538461538462e-05,
|
| 500 |
+
"loss": 3.1517,
|
| 501 |
+
"step": 492
|
| 502 |
+
},
|
| 503 |
+
{
|
| 504 |
+
"epoch": 1.6,
|
| 505 |
+
"learning_rate": 4.201923076923077e-05,
|
| 506 |
+
"loss": 3.2236,
|
| 507 |
+
"step": 498
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 1.62,
|
| 511 |
+
"learning_rate": 4.192307692307693e-05,
|
| 512 |
+
"loss": 3.1939,
|
| 513 |
+
"step": 504
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 1.63,
|
| 517 |
+
"learning_rate": 4.182692307692308e-05,
|
| 518 |
+
"loss": 3.1143,
|
| 519 |
+
"step": 510
|
| 520 |
+
},
|
| 521 |
+
{
|
| 522 |
+
"epoch": 1.65,
|
| 523 |
+
"learning_rate": 4.173076923076923e-05,
|
| 524 |
+
"loss": 3.2261,
|
| 525 |
+
"step": 516
|
| 526 |
+
},
|
| 527 |
+
{
|
| 528 |
+
"epoch": 1.67,
|
| 529 |
+
"learning_rate": 4.163461538461539e-05,
|
| 530 |
+
"loss": 3.2095,
|
| 531 |
+
"step": 522
|
| 532 |
+
},
|
| 533 |
+
{
|
| 534 |
+
"epoch": 1.69,
|
| 535 |
+
"learning_rate": 4.1538461538461544e-05,
|
| 536 |
+
"loss": 3.2498,
|
| 537 |
+
"step": 528
|
| 538 |
+
},
|
| 539 |
+
{
|
| 540 |
+
"epoch": 1.71,
|
| 541 |
+
"learning_rate": 4.1442307692307696e-05,
|
| 542 |
+
"loss": 3.1493,
|
| 543 |
+
"step": 534
|
| 544 |
+
},
|
| 545 |
+
{
|
| 546 |
+
"epoch": 1.73,
|
| 547 |
+
"learning_rate": 4.134615384615385e-05,
|
| 548 |
+
"loss": 3.0919,
|
| 549 |
+
"step": 540
|
| 550 |
+
},
|
| 551 |
+
{
|
| 552 |
+
"epoch": 1.75,
|
| 553 |
+
"learning_rate": 4.125e-05,
|
| 554 |
+
"loss": 3.2617,
|
| 555 |
+
"step": 546
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 1.77,
|
| 559 |
+
"learning_rate": 4.115384615384615e-05,
|
| 560 |
+
"loss": 3.1447,
|
| 561 |
+
"step": 552
|
| 562 |
+
},
|
| 563 |
+
{
|
| 564 |
+
"epoch": 1.79,
|
| 565 |
+
"learning_rate": 4.105769230769231e-05,
|
| 566 |
+
"loss": 3.2025,
|
| 567 |
+
"step": 558
|
| 568 |
+
},
|
| 569 |
+
{
|
| 570 |
+
"epoch": 1.81,
|
| 571 |
+
"learning_rate": 4.096153846153846e-05,
|
| 572 |
+
"loss": 3.1829,
|
| 573 |
+
"step": 564
|
| 574 |
+
},
|
| 575 |
+
{
|
| 576 |
+
"epoch": 1.83,
|
| 577 |
+
"learning_rate": 4.0865384615384615e-05,
|
| 578 |
+
"loss": 3.2216,
|
| 579 |
+
"step": 570
|
| 580 |
+
},
|
| 581 |
+
{
|
| 582 |
+
"epoch": 1.85,
|
| 583 |
+
"learning_rate": 4.0769230769230773e-05,
|
| 584 |
+
"loss": 3.1391,
|
| 585 |
+
"step": 576
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 1.87,
|
| 589 |
+
"learning_rate": 4.0673076923076926e-05,
|
| 590 |
+
"loss": 3.1344,
|
| 591 |
+
"step": 582
|
| 592 |
+
},
|
| 593 |
+
{
|
| 594 |
+
"epoch": 1.88,
|
| 595 |
+
"learning_rate": 4.057692307692308e-05,
|
| 596 |
+
"loss": 3.1699,
|
| 597 |
+
"step": 588
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 1.9,
|
| 601 |
+
"learning_rate": 4.0480769230769236e-05,
|
| 602 |
+
"loss": 3.0138,
|
| 603 |
+
"step": 594
|
| 604 |
+
},
|
| 605 |
+
{
|
| 606 |
+
"epoch": 1.92,
|
| 607 |
+
"learning_rate": 4.038461538461539e-05,
|
| 608 |
+
"loss": 3.1157,
|
| 609 |
+
"step": 600
|
| 610 |
+
},
|
| 611 |
+
{
|
| 612 |
+
"epoch": 1.94,
|
| 613 |
+
"learning_rate": 4.028846153846154e-05,
|
| 614 |
+
"loss": 3.1173,
|
| 615 |
+
"step": 606
|
| 616 |
+
},
|
| 617 |
+
{
|
| 618 |
+
"epoch": 1.96,
|
| 619 |
+
"learning_rate": 4.019230769230769e-05,
|
| 620 |
+
"loss": 3.2019,
|
| 621 |
+
"step": 612
|
| 622 |
+
},
|
| 623 |
+
{
|
| 624 |
+
"epoch": 1.98,
|
| 625 |
+
"learning_rate": 4.009615384615385e-05,
|
| 626 |
+
"loss": 3.0801,
|
| 627 |
+
"step": 618
|
| 628 |
+
},
|
| 629 |
+
{
|
| 630 |
+
"epoch": 2.0,
|
| 631 |
+
"learning_rate": 4e-05,
|
| 632 |
+
"loss": 3.0888,
|
| 633 |
+
"step": 624
|
| 634 |
+
},
|
| 635 |
+
{
|
| 636 |
+
"epoch": 2.02,
|
| 637 |
+
"learning_rate": 3.9903846153846155e-05,
|
| 638 |
+
"loss": 3.0332,
|
| 639 |
+
"step": 630
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 2.04,
|
| 643 |
+
"learning_rate": 3.980769230769231e-05,
|
| 644 |
+
"loss": 2.9418,
|
| 645 |
+
"step": 636
|
| 646 |
+
},
|
| 647 |
+
{
|
| 648 |
+
"epoch": 2.06,
|
| 649 |
+
"learning_rate": 3.971153846153846e-05,
|
| 650 |
+
"loss": 2.933,
|
| 651 |
+
"step": 642
|
| 652 |
+
},
|
| 653 |
+
{
|
| 654 |
+
"epoch": 2.08,
|
| 655 |
+
"learning_rate": 3.961538461538462e-05,
|
| 656 |
+
"loss": 3.0467,
|
| 657 |
+
"step": 648
|
| 658 |
+
},
|
| 659 |
+
{
|
| 660 |
+
"epoch": 2.1,
|
| 661 |
+
"learning_rate": 3.951923076923077e-05,
|
| 662 |
+
"loss": 3.0991,
|
| 663 |
+
"step": 654
|
| 664 |
+
},
|
| 665 |
+
{
|
| 666 |
+
"epoch": 2.12,
|
| 667 |
+
"learning_rate": 3.942307692307692e-05,
|
| 668 |
+
"loss": 2.913,
|
| 669 |
+
"step": 660
|
| 670 |
+
},
|
| 671 |
+
{
|
| 672 |
+
"epoch": 2.13,
|
| 673 |
+
"learning_rate": 3.932692307692308e-05,
|
| 674 |
+
"loss": 3.0531,
|
| 675 |
+
"step": 666
|
| 676 |
+
},
|
| 677 |
+
{
|
| 678 |
+
"epoch": 2.15,
|
| 679 |
+
"learning_rate": 3.923076923076923e-05,
|
| 680 |
+
"loss": 2.9838,
|
| 681 |
+
"step": 672
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 2.17,
|
| 685 |
+
"learning_rate": 3.9134615384615385e-05,
|
| 686 |
+
"loss": 2.9406,
|
| 687 |
+
"step": 678
|
| 688 |
+
},
|
| 689 |
+
{
|
| 690 |
+
"epoch": 2.19,
|
| 691 |
+
"learning_rate": 3.903846153846154e-05,
|
| 692 |
+
"loss": 3.046,
|
| 693 |
+
"step": 684
|
| 694 |
+
},
|
| 695 |
+
{
|
| 696 |
+
"epoch": 2.21,
|
| 697 |
+
"learning_rate": 3.8942307692307696e-05,
|
| 698 |
+
"loss": 3.006,
|
| 699 |
+
"step": 690
|
| 700 |
+
},
|
| 701 |
+
{
|
| 702 |
+
"epoch": 2.23,
|
| 703 |
+
"learning_rate": 3.884615384615385e-05,
|
| 704 |
+
"loss": 2.9774,
|
| 705 |
+
"step": 696
|
| 706 |
+
},
|
| 707 |
+
{
|
| 708 |
+
"epoch": 2.25,
|
| 709 |
+
"learning_rate": 3.875e-05,
|
| 710 |
+
"loss": 2.9937,
|
| 711 |
+
"step": 702
|
| 712 |
+
},
|
| 713 |
+
{
|
| 714 |
+
"epoch": 2.27,
|
| 715 |
+
"learning_rate": 3.865384615384616e-05,
|
| 716 |
+
"loss": 2.9737,
|
| 717 |
+
"step": 708
|
| 718 |
+
},
|
| 719 |
+
{
|
| 720 |
+
"epoch": 2.29,
|
| 721 |
+
"learning_rate": 3.855769230769231e-05,
|
| 722 |
+
"loss": 2.9772,
|
| 723 |
+
"step": 714
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 2.31,
|
| 727 |
+
"learning_rate": 3.846153846153846e-05,
|
| 728 |
+
"loss": 2.8405,
|
| 729 |
+
"step": 720
|
| 730 |
+
},
|
| 731 |
+
{
|
| 732 |
+
"epoch": 2.33,
|
| 733 |
+
"learning_rate": 3.836538461538462e-05,
|
| 734 |
+
"loss": 3.0856,
|
| 735 |
+
"step": 726
|
| 736 |
+
},
|
| 737 |
+
{
|
| 738 |
+
"epoch": 2.35,
|
| 739 |
+
"learning_rate": 3.826923076923077e-05,
|
| 740 |
+
"loss": 2.9947,
|
| 741 |
+
"step": 732
|
| 742 |
+
},
|
| 743 |
+
{
|
| 744 |
+
"epoch": 2.37,
|
| 745 |
+
"learning_rate": 3.8173076923076926e-05,
|
| 746 |
+
"loss": 2.9251,
|
| 747 |
+
"step": 738
|
| 748 |
+
},
|
| 749 |
+
{
|
| 750 |
+
"epoch": 2.38,
|
| 751 |
+
"learning_rate": 3.807692307692308e-05,
|
| 752 |
+
"loss": 2.9613,
|
| 753 |
+
"step": 744
|
| 754 |
+
},
|
| 755 |
+
{
|
| 756 |
+
"epoch": 2.4,
|
| 757 |
+
"learning_rate": 3.798076923076923e-05,
|
| 758 |
+
"loss": 2.9402,
|
| 759 |
+
"step": 750
|
| 760 |
+
},
|
| 761 |
+
{
|
| 762 |
+
"epoch": 2.42,
|
| 763 |
+
"learning_rate": 3.788461538461538e-05,
|
| 764 |
+
"loss": 2.8823,
|
| 765 |
+
"step": 756
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 2.44,
|
| 769 |
+
"learning_rate": 3.778846153846154e-05,
|
| 770 |
+
"loss": 2.8906,
|
| 771 |
+
"step": 762
|
| 772 |
+
},
|
| 773 |
+
{
|
| 774 |
+
"epoch": 2.46,
|
| 775 |
+
"learning_rate": 3.769230769230769e-05,
|
| 776 |
+
"loss": 2.9475,
|
| 777 |
+
"step": 768
|
| 778 |
+
},
|
| 779 |
+
{
|
| 780 |
+
"epoch": 2.48,
|
| 781 |
+
"learning_rate": 3.7596153846153845e-05,
|
| 782 |
+
"loss": 2.9999,
|
| 783 |
+
"step": 774
|
| 784 |
+
},
|
| 785 |
+
{
|
| 786 |
+
"epoch": 2.5,
|
| 787 |
+
"learning_rate": 3.7500000000000003e-05,
|
| 788 |
+
"loss": 2.9503,
|
| 789 |
+
"step": 780
|
| 790 |
+
},
|
| 791 |
+
{
|
| 792 |
+
"epoch": 2.52,
|
| 793 |
+
"learning_rate": 3.7403846153846156e-05,
|
| 794 |
+
"loss": 3.0112,
|
| 795 |
+
"step": 786
|
| 796 |
+
},
|
| 797 |
+
{
|
| 798 |
+
"epoch": 2.54,
|
| 799 |
+
"learning_rate": 3.730769230769231e-05,
|
| 800 |
+
"loss": 3.0096,
|
| 801 |
+
"step": 792
|
| 802 |
+
},
|
| 803 |
+
{
|
| 804 |
+
"epoch": 2.56,
|
| 805 |
+
"learning_rate": 3.7211538461538466e-05,
|
| 806 |
+
"loss": 2.9804,
|
| 807 |
+
"step": 798
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 2.58,
|
| 811 |
+
"learning_rate": 3.711538461538462e-05,
|
| 812 |
+
"loss": 2.9203,
|
| 813 |
+
"step": 804
|
| 814 |
+
},
|
| 815 |
+
{
|
| 816 |
+
"epoch": 2.6,
|
| 817 |
+
"learning_rate": 3.701923076923077e-05,
|
| 818 |
+
"loss": 2.939,
|
| 819 |
+
"step": 810
|
| 820 |
+
},
|
| 821 |
+
{
|
| 822 |
+
"epoch": 2.62,
|
| 823 |
+
"learning_rate": 3.692307692307693e-05,
|
| 824 |
+
"loss": 2.8455,
|
| 825 |
+
"step": 816
|
| 826 |
+
},
|
| 827 |
+
{
|
| 828 |
+
"epoch": 2.63,
|
| 829 |
+
"learning_rate": 3.682692307692308e-05,
|
| 830 |
+
"loss": 2.9651,
|
| 831 |
+
"step": 822
|
| 832 |
+
},
|
| 833 |
+
{
|
| 834 |
+
"epoch": 2.65,
|
| 835 |
+
"learning_rate": 3.673076923076923e-05,
|
| 836 |
+
"loss": 2.9528,
|
| 837 |
+
"step": 828
|
| 838 |
+
},
|
| 839 |
+
{
|
| 840 |
+
"epoch": 2.67,
|
| 841 |
+
"learning_rate": 3.6634615384615385e-05,
|
| 842 |
+
"loss": 2.8042,
|
| 843 |
+
"step": 834
|
| 844 |
+
},
|
| 845 |
+
{
|
| 846 |
+
"epoch": 2.69,
|
| 847 |
+
"learning_rate": 3.653846153846154e-05,
|
| 848 |
+
"loss": 2.8311,
|
| 849 |
+
"step": 840
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 2.71,
|
| 853 |
+
"learning_rate": 3.644230769230769e-05,
|
| 854 |
+
"loss": 2.8888,
|
| 855 |
+
"step": 846
|
| 856 |
+
},
|
| 857 |
+
{
|
| 858 |
+
"epoch": 2.73,
|
| 859 |
+
"learning_rate": 3.634615384615385e-05,
|
| 860 |
+
"loss": 2.9151,
|
| 861 |
+
"step": 852
|
| 862 |
+
},
|
| 863 |
+
{
|
| 864 |
+
"epoch": 2.75,
|
| 865 |
+
"learning_rate": 3.625e-05,
|
| 866 |
+
"loss": 2.9463,
|
| 867 |
+
"step": 858
|
| 868 |
+
},
|
| 869 |
+
{
|
| 870 |
+
"epoch": 2.77,
|
| 871 |
+
"learning_rate": 3.615384615384615e-05,
|
| 872 |
+
"loss": 2.957,
|
| 873 |
+
"step": 864
|
| 874 |
+
},
|
| 875 |
+
{
|
| 876 |
+
"epoch": 2.79,
|
| 877 |
+
"learning_rate": 3.605769230769231e-05,
|
| 878 |
+
"loss": 2.9473,
|
| 879 |
+
"step": 870
|
| 880 |
+
},
|
| 881 |
+
{
|
| 882 |
+
"epoch": 2.81,
|
| 883 |
+
"learning_rate": 3.596153846153846e-05,
|
| 884 |
+
"loss": 2.9994,
|
| 885 |
+
"step": 876
|
| 886 |
+
},
|
| 887 |
+
{
|
| 888 |
+
"epoch": 2.83,
|
| 889 |
+
"learning_rate": 3.5865384615384615e-05,
|
| 890 |
+
"loss": 3.0486,
|
| 891 |
+
"step": 882
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 2.85,
|
| 895 |
+
"learning_rate": 3.5769230769230774e-05,
|
| 896 |
+
"loss": 2.9487,
|
| 897 |
+
"step": 888
|
| 898 |
+
},
|
| 899 |
+
{
|
| 900 |
+
"epoch": 2.87,
|
| 901 |
+
"learning_rate": 3.5673076923076926e-05,
|
| 902 |
+
"loss": 3.0173,
|
| 903 |
+
"step": 894
|
| 904 |
+
},
|
| 905 |
+
{
|
| 906 |
+
"epoch": 2.88,
|
| 907 |
+
"learning_rate": 3.557692307692308e-05,
|
| 908 |
+
"loss": 2.8656,
|
| 909 |
+
"step": 900
|
| 910 |
+
},
|
| 911 |
+
{
|
| 912 |
+
"epoch": 2.9,
|
| 913 |
+
"learning_rate": 3.548076923076924e-05,
|
| 914 |
+
"loss": 2.8834,
|
| 915 |
+
"step": 906
|
| 916 |
+
},
|
| 917 |
+
{
|
| 918 |
+
"epoch": 2.92,
|
| 919 |
+
"learning_rate": 3.538461538461539e-05,
|
| 920 |
+
"loss": 2.9829,
|
| 921 |
+
"step": 912
|
| 922 |
+
},
|
| 923 |
+
{
|
| 924 |
+
"epoch": 2.94,
|
| 925 |
+
"learning_rate": 3.528846153846154e-05,
|
| 926 |
+
"loss": 3.001,
|
| 927 |
+
"step": 918
|
| 928 |
+
},
|
| 929 |
+
{
|
| 930 |
+
"epoch": 2.96,
|
| 931 |
+
"learning_rate": 3.51923076923077e-05,
|
| 932 |
+
"loss": 2.9618,
|
| 933 |
+
"step": 924
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 2.98,
|
| 937 |
+
"learning_rate": 3.5096153846153845e-05,
|
| 938 |
+
"loss": 2.7964,
|
| 939 |
+
"step": 930
|
| 940 |
+
},
|
| 941 |
+
{
|
| 942 |
+
"epoch": 3.0,
|
| 943 |
+
"learning_rate": 3.5e-05,
|
| 944 |
+
"loss": 2.89,
|
| 945 |
+
"step": 936
|
| 946 |
+
},
|
| 947 |
+
{
|
| 948 |
+
"epoch": 3.02,
|
| 949 |
+
"learning_rate": 3.4903846153846156e-05,
|
| 950 |
+
"loss": 2.7191,
|
| 951 |
+
"step": 942
|
| 952 |
+
},
|
| 953 |
+
{
|
| 954 |
+
"epoch": 3.04,
|
| 955 |
+
"learning_rate": 3.480769230769231e-05,
|
| 956 |
+
"loss": 2.7875,
|
| 957 |
+
"step": 948
|
| 958 |
+
},
|
| 959 |
+
{
|
| 960 |
+
"epoch": 3.06,
|
| 961 |
+
"learning_rate": 3.471153846153846e-05,
|
| 962 |
+
"loss": 2.8498,
|
| 963 |
+
"step": 954
|
| 964 |
+
},
|
| 965 |
+
{
|
| 966 |
+
"epoch": 3.08,
|
| 967 |
+
"learning_rate": 3.461538461538462e-05,
|
| 968 |
+
"loss": 2.7858,
|
| 969 |
+
"step": 960
|
| 970 |
+
},
|
| 971 |
+
{
|
| 972 |
+
"epoch": 3.1,
|
| 973 |
+
"learning_rate": 3.451923076923077e-05,
|
| 974 |
+
"loss": 2.7435,
|
| 975 |
+
"step": 966
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 3.12,
|
| 979 |
+
"learning_rate": 3.442307692307692e-05,
|
| 980 |
+
"loss": 2.8449,
|
| 981 |
+
"step": 972
|
| 982 |
+
},
|
| 983 |
+
{
|
| 984 |
+
"epoch": 3.13,
|
| 985 |
+
"learning_rate": 3.432692307692308e-05,
|
| 986 |
+
"loss": 2.7817,
|
| 987 |
+
"step": 978
|
| 988 |
+
},
|
| 989 |
+
{
|
| 990 |
+
"epoch": 3.15,
|
| 991 |
+
"learning_rate": 3.4230769230769234e-05,
|
| 992 |
+
"loss": 2.8797,
|
| 993 |
+
"step": 984
|
| 994 |
+
},
|
| 995 |
+
{
|
| 996 |
+
"epoch": 3.17,
|
| 997 |
+
"learning_rate": 3.4134615384615386e-05,
|
| 998 |
+
"loss": 2.7719,
|
| 999 |
+
"step": 990
|
| 1000 |
+
},
|
| 1001 |
+
{
|
| 1002 |
+
"epoch": 3.19,
|
| 1003 |
+
"learning_rate": 3.4038461538461544e-05,
|
| 1004 |
+
"loss": 2.7919,
|
| 1005 |
+
"step": 996
|
| 1006 |
+
},
|
| 1007 |
+
{
|
| 1008 |
+
"epoch": 3.21,
|
| 1009 |
+
"learning_rate": 3.3942307692307696e-05,
|
| 1010 |
+
"loss": 2.8072,
|
| 1011 |
+
"step": 1002
|
| 1012 |
+
},
|
| 1013 |
+
{
|
| 1014 |
+
"epoch": 3.23,
|
| 1015 |
+
"learning_rate": 3.384615384615385e-05,
|
| 1016 |
+
"loss": 2.7281,
|
| 1017 |
+
"step": 1008
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 3.25,
|
| 1021 |
+
"learning_rate": 3.375000000000001e-05,
|
| 1022 |
+
"loss": 2.8431,
|
| 1023 |
+
"step": 1014
|
| 1024 |
+
},
|
| 1025 |
+
{
|
| 1026 |
+
"epoch": 3.27,
|
| 1027 |
+
"learning_rate": 3.365384615384616e-05,
|
| 1028 |
+
"loss": 2.7154,
|
| 1029 |
+
"step": 1020
|
| 1030 |
+
},
|
| 1031 |
+
{
|
| 1032 |
+
"epoch": 3.29,
|
| 1033 |
+
"learning_rate": 3.3557692307692304e-05,
|
| 1034 |
+
"loss": 2.83,
|
| 1035 |
+
"step": 1026
|
| 1036 |
+
},
|
| 1037 |
+
{
|
| 1038 |
+
"epoch": 3.31,
|
| 1039 |
+
"learning_rate": 3.346153846153846e-05,
|
| 1040 |
+
"loss": 2.8434,
|
| 1041 |
+
"step": 1032
|
| 1042 |
+
},
|
| 1043 |
+
{
|
| 1044 |
+
"epoch": 3.33,
|
| 1045 |
+
"learning_rate": 3.3365384615384615e-05,
|
| 1046 |
+
"loss": 2.8399,
|
| 1047 |
+
"step": 1038
|
| 1048 |
+
},
|
| 1049 |
+
{
|
| 1050 |
+
"epoch": 3.35,
|
| 1051 |
+
"learning_rate": 3.326923076923077e-05,
|
| 1052 |
+
"loss": 2.7072,
|
| 1053 |
+
"step": 1044
|
| 1054 |
+
},
|
| 1055 |
+
{
|
| 1056 |
+
"epoch": 3.37,
|
| 1057 |
+
"learning_rate": 3.3173076923076926e-05,
|
| 1058 |
+
"loss": 2.773,
|
| 1059 |
+
"step": 1050
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 3.38,
|
| 1063 |
+
"learning_rate": 3.307692307692308e-05,
|
| 1064 |
+
"loss": 2.822,
|
| 1065 |
+
"step": 1056
|
| 1066 |
+
},
|
| 1067 |
+
{
|
| 1068 |
+
"epoch": 3.4,
|
| 1069 |
+
"learning_rate": 3.298076923076923e-05,
|
| 1070 |
+
"loss": 2.9056,
|
| 1071 |
+
"step": 1062
|
| 1072 |
+
},
|
| 1073 |
+
{
|
| 1074 |
+
"epoch": 3.42,
|
| 1075 |
+
"learning_rate": 3.288461538461539e-05,
|
| 1076 |
+
"loss": 2.749,
|
| 1077 |
+
"step": 1068
|
| 1078 |
+
},
|
| 1079 |
+
{
|
| 1080 |
+
"epoch": 3.44,
|
| 1081 |
+
"learning_rate": 3.278846153846154e-05,
|
| 1082 |
+
"loss": 2.7579,
|
| 1083 |
+
"step": 1074
|
| 1084 |
+
},
|
| 1085 |
+
{
|
| 1086 |
+
"epoch": 3.46,
|
| 1087 |
+
"learning_rate": 3.269230769230769e-05,
|
| 1088 |
+
"loss": 2.7444,
|
| 1089 |
+
"step": 1080
|
| 1090 |
+
},
|
| 1091 |
+
{
|
| 1092 |
+
"epoch": 3.48,
|
| 1093 |
+
"learning_rate": 3.2596153846153845e-05,
|
| 1094 |
+
"loss": 2.818,
|
| 1095 |
+
"step": 1086
|
| 1096 |
+
},
|
| 1097 |
+
{
|
| 1098 |
+
"epoch": 3.5,
|
| 1099 |
+
"learning_rate": 3.2500000000000004e-05,
|
| 1100 |
+
"loss": 2.6285,
|
| 1101 |
+
"step": 1092
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 3.52,
|
| 1105 |
+
"learning_rate": 3.2403846153846156e-05,
|
| 1106 |
+
"loss": 2.6473,
|
| 1107 |
+
"step": 1098
|
| 1108 |
+
},
|
| 1109 |
+
{
|
| 1110 |
+
"epoch": 3.54,
|
| 1111 |
+
"learning_rate": 3.230769230769231e-05,
|
| 1112 |
+
"loss": 2.7093,
|
| 1113 |
+
"step": 1104
|
| 1114 |
+
},
|
| 1115 |
+
{
|
| 1116 |
+
"epoch": 3.56,
|
| 1117 |
+
"learning_rate": 3.221153846153847e-05,
|
| 1118 |
+
"loss": 2.732,
|
| 1119 |
+
"step": 1110
|
| 1120 |
+
},
|
| 1121 |
+
{
|
| 1122 |
+
"epoch": 3.58,
|
| 1123 |
+
"learning_rate": 3.211538461538462e-05,
|
| 1124 |
+
"loss": 2.7064,
|
| 1125 |
+
"step": 1116
|
| 1126 |
+
},
|
| 1127 |
+
{
|
| 1128 |
+
"epoch": 3.6,
|
| 1129 |
+
"learning_rate": 3.201923076923077e-05,
|
| 1130 |
+
"loss": 2.9072,
|
| 1131 |
+
"step": 1122
|
| 1132 |
+
},
|
| 1133 |
+
{
|
| 1134 |
+
"epoch": 3.62,
|
| 1135 |
+
"learning_rate": 3.192307692307692e-05,
|
| 1136 |
+
"loss": 2.7544,
|
| 1137 |
+
"step": 1128
|
| 1138 |
+
},
|
| 1139 |
+
{
|
| 1140 |
+
"epoch": 3.63,
|
| 1141 |
+
"learning_rate": 3.1826923076923075e-05,
|
| 1142 |
+
"loss": 2.738,
|
| 1143 |
+
"step": 1134
|
| 1144 |
+
},
|
| 1145 |
+
{
|
| 1146 |
+
"epoch": 3.65,
|
| 1147 |
+
"learning_rate": 3.1730769230769234e-05,
|
| 1148 |
+
"loss": 2.6721,
|
| 1149 |
+
"step": 1140
|
| 1150 |
+
},
|
| 1151 |
+
{
|
| 1152 |
+
"epoch": 3.67,
|
| 1153 |
+
"learning_rate": 3.1634615384615386e-05,
|
| 1154 |
+
"loss": 2.8218,
|
| 1155 |
+
"step": 1146
|
| 1156 |
+
},
|
| 1157 |
+
{
|
| 1158 |
+
"epoch": 3.69,
|
| 1159 |
+
"learning_rate": 3.153846153846154e-05,
|
| 1160 |
+
"loss": 2.7553,
|
| 1161 |
+
"step": 1152
|
| 1162 |
+
},
|
| 1163 |
+
{
|
| 1164 |
+
"epoch": 3.71,
|
| 1165 |
+
"learning_rate": 3.144230769230769e-05,
|
| 1166 |
+
"loss": 2.8186,
|
| 1167 |
+
"step": 1158
|
| 1168 |
+
},
|
| 1169 |
+
{
|
| 1170 |
+
"epoch": 3.73,
|
| 1171 |
+
"learning_rate": 3.134615384615385e-05,
|
| 1172 |
+
"loss": 2.7511,
|
| 1173 |
+
"step": 1164
|
| 1174 |
+
},
|
| 1175 |
+
{
|
| 1176 |
+
"epoch": 3.75,
|
| 1177 |
+
"learning_rate": 3.125e-05,
|
| 1178 |
+
"loss": 2.6013,
|
| 1179 |
+
"step": 1170
|
| 1180 |
+
},
|
| 1181 |
+
{
|
| 1182 |
+
"epoch": 3.77,
|
| 1183 |
+
"learning_rate": 3.115384615384615e-05,
|
| 1184 |
+
"loss": 2.7191,
|
| 1185 |
+
"step": 1176
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 3.79,
|
| 1189 |
+
"learning_rate": 3.105769230769231e-05,
|
| 1190 |
+
"loss": 2.7695,
|
| 1191 |
+
"step": 1182
|
| 1192 |
+
},
|
| 1193 |
+
{
|
| 1194 |
+
"epoch": 3.81,
|
| 1195 |
+
"learning_rate": 3.0961538461538464e-05,
|
| 1196 |
+
"loss": 2.738,
|
| 1197 |
+
"step": 1188
|
| 1198 |
+
},
|
| 1199 |
+
{
|
| 1200 |
+
"epoch": 3.83,
|
| 1201 |
+
"learning_rate": 3.0865384615384616e-05,
|
| 1202 |
+
"loss": 2.7202,
|
| 1203 |
+
"step": 1194
|
| 1204 |
+
},
|
| 1205 |
+
{
|
| 1206 |
+
"epoch": 3.85,
|
| 1207 |
+
"learning_rate": 3.0769230769230774e-05,
|
| 1208 |
+
"loss": 2.6429,
|
| 1209 |
+
"step": 1200
|
| 1210 |
+
},
|
| 1211 |
+
{
|
| 1212 |
+
"epoch": 3.87,
|
| 1213 |
+
"learning_rate": 3.0673076923076926e-05,
|
| 1214 |
+
"loss": 2.6742,
|
| 1215 |
+
"step": 1206
|
| 1216 |
+
},
|
| 1217 |
+
{
|
| 1218 |
+
"epoch": 3.88,
|
| 1219 |
+
"learning_rate": 3.057692307692308e-05,
|
| 1220 |
+
"loss": 2.7048,
|
| 1221 |
+
"step": 1212
|
| 1222 |
+
},
|
| 1223 |
+
{
|
| 1224 |
+
"epoch": 3.9,
|
| 1225 |
+
"learning_rate": 3.0480769230769234e-05,
|
| 1226 |
+
"loss": 2.7016,
|
| 1227 |
+
"step": 1218
|
| 1228 |
+
},
|
| 1229 |
+
{
|
| 1230 |
+
"epoch": 3.92,
|
| 1231 |
+
"learning_rate": 3.0384615384615382e-05,
|
| 1232 |
+
"loss": 2.7375,
|
| 1233 |
+
"step": 1224
|
| 1234 |
+
},
|
| 1235 |
+
{
|
| 1236 |
+
"epoch": 3.94,
|
| 1237 |
+
"learning_rate": 3.0288461538461538e-05,
|
| 1238 |
+
"loss": 2.7926,
|
| 1239 |
+
"step": 1230
|
| 1240 |
+
},
|
| 1241 |
+
{
|
| 1242 |
+
"epoch": 3.96,
|
| 1243 |
+
"learning_rate": 3.0192307692307693e-05,
|
| 1244 |
+
"loss": 2.7038,
|
| 1245 |
+
"step": 1236
|
| 1246 |
+
},
|
| 1247 |
+
{
|
| 1248 |
+
"epoch": 3.98,
|
| 1249 |
+
"learning_rate": 3.0096153846153845e-05,
|
| 1250 |
+
"loss": 2.8294,
|
| 1251 |
+
"step": 1242
|
| 1252 |
+
},
|
| 1253 |
+
{
|
| 1254 |
+
"epoch": 4.0,
|
| 1255 |
+
"learning_rate": 3e-05,
|
| 1256 |
+
"loss": 2.7543,
|
| 1257 |
+
"step": 1248
|
| 1258 |
+
},
|
| 1259 |
+
{
|
| 1260 |
+
"epoch": 4.02,
|
| 1261 |
+
"learning_rate": 2.9903846153846156e-05,
|
| 1262 |
+
"loss": 2.6022,
|
| 1263 |
+
"step": 1254
|
| 1264 |
+
},
|
| 1265 |
+
{
|
| 1266 |
+
"epoch": 4.04,
|
| 1267 |
+
"learning_rate": 2.9807692307692308e-05,
|
| 1268 |
+
"loss": 2.6609,
|
| 1269 |
+
"step": 1260
|
| 1270 |
+
},
|
| 1271 |
+
{
|
| 1272 |
+
"epoch": 4.06,
|
| 1273 |
+
"learning_rate": 2.9711538461538464e-05,
|
| 1274 |
+
"loss": 2.6091,
|
| 1275 |
+
"step": 1266
|
| 1276 |
+
},
|
| 1277 |
+
{
|
| 1278 |
+
"epoch": 4.08,
|
| 1279 |
+
"learning_rate": 2.9615384615384616e-05,
|
| 1280 |
+
"loss": 2.5386,
|
| 1281 |
+
"step": 1272
|
| 1282 |
+
},
|
| 1283 |
+
{
|
| 1284 |
+
"epoch": 4.1,
|
| 1285 |
+
"learning_rate": 2.951923076923077e-05,
|
| 1286 |
+
"loss": 2.6655,
|
| 1287 |
+
"step": 1278
|
| 1288 |
+
},
|
| 1289 |
+
{
|
| 1290 |
+
"epoch": 4.12,
|
| 1291 |
+
"learning_rate": 2.9423076923076926e-05,
|
| 1292 |
+
"loss": 2.6023,
|
| 1293 |
+
"step": 1284
|
| 1294 |
+
},
|
| 1295 |
+
{
|
| 1296 |
+
"epoch": 4.13,
|
| 1297 |
+
"learning_rate": 2.932692307692308e-05,
|
| 1298 |
+
"loss": 2.6491,
|
| 1299 |
+
"step": 1290
|
| 1300 |
+
},
|
| 1301 |
+
{
|
| 1302 |
+
"epoch": 4.15,
|
| 1303 |
+
"learning_rate": 2.9230769230769234e-05,
|
| 1304 |
+
"loss": 2.6548,
|
| 1305 |
+
"step": 1296
|
| 1306 |
+
},
|
| 1307 |
+
{
|
| 1308 |
+
"epoch": 4.17,
|
| 1309 |
+
"learning_rate": 2.913461538461539e-05,
|
| 1310 |
+
"loss": 2.584,
|
| 1311 |
+
"step": 1302
|
| 1312 |
+
},
|
| 1313 |
+
{
|
| 1314 |
+
"epoch": 4.19,
|
| 1315 |
+
"learning_rate": 2.903846153846154e-05,
|
| 1316 |
+
"loss": 2.5447,
|
| 1317 |
+
"step": 1308
|
| 1318 |
+
},
|
| 1319 |
+
{
|
| 1320 |
+
"epoch": 4.21,
|
| 1321 |
+
"learning_rate": 2.8942307692307697e-05,
|
| 1322 |
+
"loss": 2.5931,
|
| 1323 |
+
"step": 1314
|
| 1324 |
+
},
|
| 1325 |
+
{
|
| 1326 |
+
"epoch": 4.23,
|
| 1327 |
+
"learning_rate": 2.8846153846153845e-05,
|
| 1328 |
+
"loss": 2.503,
|
| 1329 |
+
"step": 1320
|
| 1330 |
+
},
|
| 1331 |
+
{
|
| 1332 |
+
"epoch": 4.25,
|
| 1333 |
+
"learning_rate": 2.8749999999999997e-05,
|
| 1334 |
+
"loss": 2.6097,
|
| 1335 |
+
"step": 1326
|
| 1336 |
+
},
|
| 1337 |
+
{
|
| 1338 |
+
"epoch": 4.27,
|
| 1339 |
+
"learning_rate": 2.8653846153846153e-05,
|
| 1340 |
+
"loss": 2.6926,
|
| 1341 |
+
"step": 1332
|
| 1342 |
+
},
|
| 1343 |
+
{
|
| 1344 |
+
"epoch": 4.29,
|
| 1345 |
+
"learning_rate": 2.855769230769231e-05,
|
| 1346 |
+
"loss": 2.5829,
|
| 1347 |
+
"step": 1338
|
| 1348 |
+
},
|
| 1349 |
+
{
|
| 1350 |
+
"epoch": 4.31,
|
| 1351 |
+
"learning_rate": 2.846153846153846e-05,
|
| 1352 |
+
"loss": 2.6712,
|
| 1353 |
+
"step": 1344
|
| 1354 |
+
},
|
| 1355 |
+
{
|
| 1356 |
+
"epoch": 4.33,
|
| 1357 |
+
"learning_rate": 2.8365384615384616e-05,
|
| 1358 |
+
"loss": 2.3988,
|
| 1359 |
+
"step": 1350
|
| 1360 |
+
},
|
| 1361 |
+
{
|
| 1362 |
+
"epoch": 4.35,
|
| 1363 |
+
"learning_rate": 2.826923076923077e-05,
|
| 1364 |
+
"loss": 2.5714,
|
| 1365 |
+
"step": 1356
|
| 1366 |
+
},
|
| 1367 |
+
{
|
| 1368 |
+
"epoch": 4.37,
|
| 1369 |
+
"learning_rate": 2.8173076923076923e-05,
|
| 1370 |
+
"loss": 2.6627,
|
| 1371 |
+
"step": 1362
|
| 1372 |
+
},
|
| 1373 |
+
{
|
| 1374 |
+
"epoch": 4.38,
|
| 1375 |
+
"learning_rate": 2.807692307692308e-05,
|
| 1376 |
+
"loss": 2.7195,
|
| 1377 |
+
"step": 1368
|
| 1378 |
+
},
|
| 1379 |
+
{
|
| 1380 |
+
"epoch": 4.4,
|
| 1381 |
+
"learning_rate": 2.7980769230769234e-05,
|
| 1382 |
+
"loss": 2.7181,
|
| 1383 |
+
"step": 1374
|
| 1384 |
+
},
|
| 1385 |
+
{
|
| 1386 |
+
"epoch": 4.42,
|
| 1387 |
+
"learning_rate": 2.7884615384615386e-05,
|
| 1388 |
+
"loss": 2.626,
|
| 1389 |
+
"step": 1380
|
| 1390 |
+
},
|
| 1391 |
+
{
|
| 1392 |
+
"epoch": 4.44,
|
| 1393 |
+
"learning_rate": 2.778846153846154e-05,
|
| 1394 |
+
"loss": 2.638,
|
| 1395 |
+
"step": 1386
|
| 1396 |
+
},
|
| 1397 |
+
{
|
| 1398 |
+
"epoch": 4.46,
|
| 1399 |
+
"learning_rate": 2.7692307692307694e-05,
|
| 1400 |
+
"loss": 2.6275,
|
| 1401 |
+
"step": 1392
|
| 1402 |
+
},
|
| 1403 |
+
{
|
| 1404 |
+
"epoch": 4.48,
|
| 1405 |
+
"learning_rate": 2.759615384615385e-05,
|
| 1406 |
+
"loss": 2.6399,
|
| 1407 |
+
"step": 1398
|
| 1408 |
+
},
|
| 1409 |
+
{
|
| 1410 |
+
"epoch": 4.5,
|
| 1411 |
+
"learning_rate": 2.7500000000000004e-05,
|
| 1412 |
+
"loss": 2.6894,
|
| 1413 |
+
"step": 1404
|
| 1414 |
+
},
|
| 1415 |
+
{
|
| 1416 |
+
"epoch": 4.52,
|
| 1417 |
+
"learning_rate": 2.7403846153846156e-05,
|
| 1418 |
+
"loss": 2.5917,
|
| 1419 |
+
"step": 1410
|
| 1420 |
+
},
|
| 1421 |
+
{
|
| 1422 |
+
"epoch": 4.54,
|
| 1423 |
+
"learning_rate": 2.7307692307692305e-05,
|
| 1424 |
+
"loss": 2.5564,
|
| 1425 |
+
"step": 1416
|
| 1426 |
+
},
|
| 1427 |
+
{
|
| 1428 |
+
"epoch": 4.56,
|
| 1429 |
+
"learning_rate": 2.721153846153846e-05,
|
| 1430 |
+
"loss": 2.6134,
|
| 1431 |
+
"step": 1422
|
| 1432 |
+
},
|
| 1433 |
+
{
|
| 1434 |
+
"epoch": 4.58,
|
| 1435 |
+
"learning_rate": 2.7115384615384616e-05,
|
| 1436 |
+
"loss": 2.6375,
|
| 1437 |
+
"step": 1428
|
| 1438 |
+
},
|
| 1439 |
+
{
|
| 1440 |
+
"epoch": 4.6,
|
| 1441 |
+
"learning_rate": 2.7019230769230768e-05,
|
| 1442 |
+
"loss": 2.6301,
|
| 1443 |
+
"step": 1434
|
| 1444 |
+
},
|
| 1445 |
+
{
|
| 1446 |
+
"epoch": 4.62,
|
| 1447 |
+
"learning_rate": 2.6923076923076923e-05,
|
| 1448 |
+
"loss": 2.5679,
|
| 1449 |
+
"step": 1440
|
| 1450 |
+
},
|
| 1451 |
+
{
|
| 1452 |
+
"epoch": 4.63,
|
| 1453 |
+
"learning_rate": 2.682692307692308e-05,
|
| 1454 |
+
"loss": 2.6704,
|
| 1455 |
+
"step": 1446
|
| 1456 |
+
},
|
| 1457 |
+
{
|
| 1458 |
+
"epoch": 4.65,
|
| 1459 |
+
"learning_rate": 2.673076923076923e-05,
|
| 1460 |
+
"loss": 2.6778,
|
| 1461 |
+
"step": 1452
|
| 1462 |
+
},
|
| 1463 |
+
{
|
| 1464 |
+
"epoch": 4.67,
|
| 1465 |
+
"learning_rate": 2.6634615384615386e-05,
|
| 1466 |
+
"loss": 2.5626,
|
| 1467 |
+
"step": 1458
|
| 1468 |
+
},
|
| 1469 |
+
{
|
| 1470 |
+
"epoch": 4.69,
|
| 1471 |
+
"learning_rate": 2.6538461538461538e-05,
|
| 1472 |
+
"loss": 2.5557,
|
| 1473 |
+
"step": 1464
|
| 1474 |
+
},
|
| 1475 |
+
{
|
| 1476 |
+
"epoch": 4.71,
|
| 1477 |
+
"learning_rate": 2.6442307692307694e-05,
|
| 1478 |
+
"loss": 2.5245,
|
| 1479 |
+
"step": 1470
|
| 1480 |
+
},
|
| 1481 |
+
{
|
| 1482 |
+
"epoch": 4.73,
|
| 1483 |
+
"learning_rate": 2.634615384615385e-05,
|
| 1484 |
+
"loss": 2.5189,
|
| 1485 |
+
"step": 1476
|
| 1486 |
+
},
|
| 1487 |
+
{
|
| 1488 |
+
"epoch": 4.75,
|
| 1489 |
+
"learning_rate": 2.625e-05,
|
| 1490 |
+
"loss": 2.4994,
|
| 1491 |
+
"step": 1482
|
| 1492 |
+
},
|
| 1493 |
+
{
|
| 1494 |
+
"epoch": 4.77,
|
| 1495 |
+
"learning_rate": 2.6153846153846157e-05,
|
| 1496 |
+
"loss": 2.472,
|
| 1497 |
+
"step": 1488
|
| 1498 |
+
},
|
| 1499 |
+
{
|
| 1500 |
+
"epoch": 4.79,
|
| 1501 |
+
"learning_rate": 2.6057692307692312e-05,
|
| 1502 |
+
"loss": 2.5451,
|
| 1503 |
+
"step": 1494
|
| 1504 |
+
},
|
| 1505 |
+
{
|
| 1506 |
+
"epoch": 4.81,
|
| 1507 |
+
"learning_rate": 2.5961538461538464e-05,
|
| 1508 |
+
"loss": 2.5559,
|
| 1509 |
+
"step": 1500
|
| 1510 |
+
},
|
| 1511 |
+
{
|
| 1512 |
+
"epoch": 4.83,
|
| 1513 |
+
"learning_rate": 2.586538461538462e-05,
|
| 1514 |
+
"loss": 2.5902,
|
| 1515 |
+
"step": 1506
|
| 1516 |
+
},
|
| 1517 |
+
{
|
| 1518 |
+
"epoch": 4.85,
|
| 1519 |
+
"learning_rate": 2.5769230769230768e-05,
|
| 1520 |
+
"loss": 2.6841,
|
| 1521 |
+
"step": 1512
|
| 1522 |
+
},
|
| 1523 |
+
{
|
| 1524 |
+
"epoch": 4.87,
|
| 1525 |
+
"learning_rate": 2.5673076923076923e-05,
|
| 1526 |
+
"loss": 2.6092,
|
| 1527 |
+
"step": 1518
|
| 1528 |
+
},
|
| 1529 |
+
{
|
| 1530 |
+
"epoch": 4.88,
|
| 1531 |
+
"learning_rate": 2.5576923076923075e-05,
|
| 1532 |
+
"loss": 2.5646,
|
| 1533 |
+
"step": 1524
|
| 1534 |
+
},
|
| 1535 |
+
{
|
| 1536 |
+
"epoch": 4.9,
|
| 1537 |
+
"learning_rate": 2.548076923076923e-05,
|
| 1538 |
+
"loss": 2.5201,
|
| 1539 |
+
"step": 1530
|
| 1540 |
+
},
|
| 1541 |
+
{
|
| 1542 |
+
"epoch": 4.92,
|
| 1543 |
+
"learning_rate": 2.5384615384615383e-05,
|
| 1544 |
+
"loss": 2.5622,
|
| 1545 |
+
"step": 1536
|
| 1546 |
+
},
|
| 1547 |
+
{
|
| 1548 |
+
"epoch": 4.94,
|
| 1549 |
+
"learning_rate": 2.528846153846154e-05,
|
| 1550 |
+
"loss": 2.6323,
|
| 1551 |
+
"step": 1542
|
| 1552 |
+
},
|
| 1553 |
+
{
|
| 1554 |
+
"epoch": 4.96,
|
| 1555 |
+
"learning_rate": 2.5192307692307694e-05,
|
| 1556 |
+
"loss": 2.7179,
|
| 1557 |
+
"step": 1548
|
| 1558 |
+
},
|
| 1559 |
+
{
|
| 1560 |
+
"epoch": 4.98,
|
| 1561 |
+
"learning_rate": 2.5096153846153846e-05,
|
| 1562 |
+
"loss": 2.6441,
|
| 1563 |
+
"step": 1554
|
| 1564 |
+
},
|
| 1565 |
+
{
|
| 1566 |
+
"epoch": 5.0,
|
| 1567 |
+
"learning_rate": 2.5e-05,
|
| 1568 |
+
"loss": 2.668,
|
| 1569 |
+
"step": 1560
|
| 1570 |
+
},
|
| 1571 |
+
{
|
| 1572 |
+
"epoch": 5.02,
|
| 1573 |
+
"learning_rate": 2.4903846153846157e-05,
|
| 1574 |
+
"loss": 2.5535,
|
| 1575 |
+
"step": 1566
|
| 1576 |
+
},
|
| 1577 |
+
{
|
| 1578 |
+
"epoch": 5.04,
|
| 1579 |
+
"learning_rate": 2.480769230769231e-05,
|
| 1580 |
+
"loss": 2.4569,
|
| 1581 |
+
"step": 1572
|
| 1582 |
+
},
|
| 1583 |
+
{
|
| 1584 |
+
"epoch": 5.06,
|
| 1585 |
+
"learning_rate": 2.4711538461538464e-05,
|
| 1586 |
+
"loss": 2.5378,
|
| 1587 |
+
"step": 1578
|
| 1588 |
+
},
|
| 1589 |
+
{
|
| 1590 |
+
"epoch": 5.08,
|
| 1591 |
+
"learning_rate": 2.461538461538462e-05,
|
| 1592 |
+
"loss": 2.537,
|
| 1593 |
+
"step": 1584
|
| 1594 |
+
},
|
| 1595 |
+
{
|
| 1596 |
+
"epoch": 5.1,
|
| 1597 |
+
"learning_rate": 2.4519230769230768e-05,
|
| 1598 |
+
"loss": 2.4913,
|
| 1599 |
+
"step": 1590
|
| 1600 |
+
},
|
| 1601 |
+
{
|
| 1602 |
+
"epoch": 5.12,
|
| 1603 |
+
"learning_rate": 2.4423076923076924e-05,
|
| 1604 |
+
"loss": 2.5156,
|
| 1605 |
+
"step": 1596
|
| 1606 |
+
},
|
| 1607 |
+
{
|
| 1608 |
+
"epoch": 5.13,
|
| 1609 |
+
"learning_rate": 2.432692307692308e-05,
|
| 1610 |
+
"loss": 2.493,
|
| 1611 |
+
"step": 1602
|
| 1612 |
+
},
|
| 1613 |
+
{
|
| 1614 |
+
"epoch": 5.15,
|
| 1615 |
+
"learning_rate": 2.423076923076923e-05,
|
| 1616 |
+
"loss": 2.5357,
|
| 1617 |
+
"step": 1608
|
| 1618 |
+
},
|
| 1619 |
+
{
|
| 1620 |
+
"epoch": 5.17,
|
| 1621 |
+
"learning_rate": 2.4134615384615386e-05,
|
| 1622 |
+
"loss": 2.4508,
|
| 1623 |
+
"step": 1614
|
| 1624 |
+
},
|
| 1625 |
+
{
|
| 1626 |
+
"epoch": 5.19,
|
| 1627 |
+
"learning_rate": 2.4038461538461542e-05,
|
| 1628 |
+
"loss": 2.5045,
|
| 1629 |
+
"step": 1620
|
| 1630 |
+
},
|
| 1631 |
+
{
|
| 1632 |
+
"epoch": 5.21,
|
| 1633 |
+
"learning_rate": 2.3942307692307694e-05,
|
| 1634 |
+
"loss": 2.4617,
|
| 1635 |
+
"step": 1626
|
| 1636 |
+
},
|
| 1637 |
+
{
|
| 1638 |
+
"epoch": 5.23,
|
| 1639 |
+
"learning_rate": 2.384615384615385e-05,
|
| 1640 |
+
"loss": 2.5392,
|
| 1641 |
+
"step": 1632
|
| 1642 |
+
},
|
| 1643 |
+
{
|
| 1644 |
+
"epoch": 5.25,
|
| 1645 |
+
"learning_rate": 2.375e-05,
|
| 1646 |
+
"loss": 2.3976,
|
| 1647 |
+
"step": 1638
|
| 1648 |
+
},
|
| 1649 |
+
{
|
| 1650 |
+
"epoch": 5.27,
|
| 1651 |
+
"learning_rate": 2.3653846153846153e-05,
|
| 1652 |
+
"loss": 2.5288,
|
| 1653 |
+
"step": 1644
|
| 1654 |
+
},
|
| 1655 |
+
{
|
| 1656 |
+
"epoch": 5.29,
|
| 1657 |
+
"learning_rate": 2.355769230769231e-05,
|
| 1658 |
+
"loss": 2.5243,
|
| 1659 |
+
"step": 1650
|
| 1660 |
+
},
|
| 1661 |
+
{
|
| 1662 |
+
"epoch": 5.31,
|
| 1663 |
+
"learning_rate": 2.3461538461538464e-05,
|
| 1664 |
+
"loss": 2.5018,
|
| 1665 |
+
"step": 1656
|
| 1666 |
+
},
|
| 1667 |
+
{
|
| 1668 |
+
"epoch": 5.33,
|
| 1669 |
+
"learning_rate": 2.3365384615384616e-05,
|
| 1670 |
+
"loss": 2.5947,
|
| 1671 |
+
"step": 1662
|
| 1672 |
+
},
|
| 1673 |
+
{
|
| 1674 |
+
"epoch": 5.35,
|
| 1675 |
+
"learning_rate": 2.326923076923077e-05,
|
| 1676 |
+
"loss": 2.5226,
|
| 1677 |
+
"step": 1668
|
| 1678 |
+
},
|
| 1679 |
+
{
|
| 1680 |
+
"epoch": 5.37,
|
| 1681 |
+
"learning_rate": 2.3173076923076924e-05,
|
| 1682 |
+
"loss": 2.4687,
|
| 1683 |
+
"step": 1674
|
| 1684 |
+
},
|
| 1685 |
+
{
|
| 1686 |
+
"epoch": 5.38,
|
| 1687 |
+
"learning_rate": 2.307692307692308e-05,
|
| 1688 |
+
"loss": 2.5297,
|
| 1689 |
+
"step": 1680
|
| 1690 |
+
},
|
| 1691 |
+
{
|
| 1692 |
+
"epoch": 5.4,
|
| 1693 |
+
"learning_rate": 2.298076923076923e-05,
|
| 1694 |
+
"loss": 2.5457,
|
| 1695 |
+
"step": 1686
|
| 1696 |
+
},
|
| 1697 |
+
{
|
| 1698 |
+
"epoch": 5.42,
|
| 1699 |
+
"learning_rate": 2.2884615384615387e-05,
|
| 1700 |
+
"loss": 2.4218,
|
| 1701 |
+
"step": 1692
|
| 1702 |
+
},
|
| 1703 |
+
{
|
| 1704 |
+
"epoch": 5.44,
|
| 1705 |
+
"learning_rate": 2.278846153846154e-05,
|
| 1706 |
+
"loss": 2.5219,
|
| 1707 |
+
"step": 1698
|
| 1708 |
+
},
|
| 1709 |
+
{
|
| 1710 |
+
"epoch": 5.46,
|
| 1711 |
+
"learning_rate": 2.2692307692307694e-05,
|
| 1712 |
+
"loss": 2.4858,
|
| 1713 |
+
"step": 1704
|
| 1714 |
+
},
|
| 1715 |
+
{
|
| 1716 |
+
"epoch": 5.48,
|
| 1717 |
+
"learning_rate": 2.2596153846153846e-05,
|
| 1718 |
+
"loss": 2.4715,
|
| 1719 |
+
"step": 1710
|
| 1720 |
+
},
|
| 1721 |
+
{
|
| 1722 |
+
"epoch": 5.5,
|
| 1723 |
+
"learning_rate": 2.25e-05,
|
| 1724 |
+
"loss": 2.3294,
|
| 1725 |
+
"step": 1716
|
| 1726 |
+
},
|
| 1727 |
+
{
|
| 1728 |
+
"epoch": 5.52,
|
| 1729 |
+
"learning_rate": 2.2403846153846157e-05,
|
| 1730 |
+
"loss": 2.5062,
|
| 1731 |
+
"step": 1722
|
| 1732 |
+
},
|
| 1733 |
+
{
|
| 1734 |
+
"epoch": 5.54,
|
| 1735 |
+
"learning_rate": 2.230769230769231e-05,
|
| 1736 |
+
"loss": 2.392,
|
| 1737 |
+
"step": 1728
|
| 1738 |
+
},
|
| 1739 |
+
{
|
| 1740 |
+
"epoch": 5.56,
|
| 1741 |
+
"learning_rate": 2.221153846153846e-05,
|
| 1742 |
+
"loss": 2.4092,
|
| 1743 |
+
"step": 1734
|
| 1744 |
+
},
|
| 1745 |
+
{
|
| 1746 |
+
"epoch": 5.58,
|
| 1747 |
+
"learning_rate": 2.2115384615384616e-05,
|
| 1748 |
+
"loss": 2.4257,
|
| 1749 |
+
"step": 1740
|
| 1750 |
+
},
|
| 1751 |
+
{
|
| 1752 |
+
"epoch": 5.6,
|
| 1753 |
+
"learning_rate": 2.201923076923077e-05,
|
| 1754 |
+
"loss": 2.4426,
|
| 1755 |
+
"step": 1746
|
| 1756 |
+
},
|
| 1757 |
+
{
|
| 1758 |
+
"epoch": 5.62,
|
| 1759 |
+
"learning_rate": 2.1923076923076924e-05,
|
| 1760 |
+
"loss": 2.4672,
|
| 1761 |
+
"step": 1752
|
| 1762 |
+
},
|
| 1763 |
+
{
|
| 1764 |
+
"epoch": 5.63,
|
| 1765 |
+
"learning_rate": 2.182692307692308e-05,
|
| 1766 |
+
"loss": 2.5468,
|
| 1767 |
+
"step": 1758
|
| 1768 |
+
},
|
| 1769 |
+
{
|
| 1770 |
+
"epoch": 5.65,
|
| 1771 |
+
"learning_rate": 2.173076923076923e-05,
|
| 1772 |
+
"loss": 2.3917,
|
| 1773 |
+
"step": 1764
|
| 1774 |
+
},
|
| 1775 |
+
{
|
| 1776 |
+
"epoch": 5.67,
|
| 1777 |
+
"learning_rate": 2.1634615384615387e-05,
|
| 1778 |
+
"loss": 2.3895,
|
| 1779 |
+
"step": 1770
|
| 1780 |
+
},
|
| 1781 |
+
{
|
| 1782 |
+
"epoch": 5.69,
|
| 1783 |
+
"learning_rate": 2.1538461538461542e-05,
|
| 1784 |
+
"loss": 2.5246,
|
| 1785 |
+
"step": 1776
|
| 1786 |
+
},
|
| 1787 |
+
{
|
| 1788 |
+
"epoch": 5.71,
|
| 1789 |
+
"learning_rate": 2.144230769230769e-05,
|
| 1790 |
+
"loss": 2.5025,
|
| 1791 |
+
"step": 1782
|
| 1792 |
+
},
|
| 1793 |
+
{
|
| 1794 |
+
"epoch": 5.73,
|
| 1795 |
+
"learning_rate": 2.1346153846153846e-05,
|
| 1796 |
+
"loss": 2.4898,
|
| 1797 |
+
"step": 1788
|
| 1798 |
+
},
|
| 1799 |
+
{
|
| 1800 |
+
"epoch": 5.75,
|
| 1801 |
+
"learning_rate": 2.125e-05,
|
| 1802 |
+
"loss": 2.5358,
|
| 1803 |
+
"step": 1794
|
| 1804 |
+
},
|
| 1805 |
+
{
|
| 1806 |
+
"epoch": 5.77,
|
| 1807 |
+
"learning_rate": 2.1153846153846154e-05,
|
| 1808 |
+
"loss": 2.5346,
|
| 1809 |
+
"step": 1800
|
| 1810 |
+
},
|
| 1811 |
+
{
|
| 1812 |
+
"epoch": 5.79,
|
| 1813 |
+
"learning_rate": 2.105769230769231e-05,
|
| 1814 |
+
"loss": 2.3057,
|
| 1815 |
+
"step": 1806
|
| 1816 |
+
},
|
| 1817 |
+
{
|
| 1818 |
+
"epoch": 5.81,
|
| 1819 |
+
"learning_rate": 2.0961538461538464e-05,
|
| 1820 |
+
"loss": 2.5392,
|
| 1821 |
+
"step": 1812
|
| 1822 |
+
},
|
| 1823 |
+
{
|
| 1824 |
+
"epoch": 5.83,
|
| 1825 |
+
"learning_rate": 2.0865384615384616e-05,
|
| 1826 |
+
"loss": 2.5815,
|
| 1827 |
+
"step": 1818
|
| 1828 |
+
},
|
| 1829 |
+
{
|
| 1830 |
+
"epoch": 5.85,
|
| 1831 |
+
"learning_rate": 2.0769230769230772e-05,
|
| 1832 |
+
"loss": 2.2718,
|
| 1833 |
+
"step": 1824
|
| 1834 |
+
},
|
| 1835 |
+
{
|
| 1836 |
+
"epoch": 5.87,
|
| 1837 |
+
"learning_rate": 2.0673076923076924e-05,
|
| 1838 |
+
"loss": 2.474,
|
| 1839 |
+
"step": 1830
|
| 1840 |
+
},
|
| 1841 |
+
{
|
| 1842 |
+
"epoch": 5.88,
|
| 1843 |
+
"learning_rate": 2.0576923076923076e-05,
|
| 1844 |
+
"loss": 2.4821,
|
| 1845 |
+
"step": 1836
|
| 1846 |
+
},
|
| 1847 |
+
{
|
| 1848 |
+
"epoch": 5.9,
|
| 1849 |
+
"learning_rate": 2.048076923076923e-05,
|
| 1850 |
+
"loss": 2.5334,
|
| 1851 |
+
"step": 1842
|
| 1852 |
+
},
|
| 1853 |
+
{
|
| 1854 |
+
"epoch": 5.92,
|
| 1855 |
+
"learning_rate": 2.0384615384615387e-05,
|
| 1856 |
+
"loss": 2.3877,
|
| 1857 |
+
"step": 1848
|
| 1858 |
+
},
|
| 1859 |
+
{
|
| 1860 |
+
"epoch": 5.94,
|
| 1861 |
+
"learning_rate": 2.028846153846154e-05,
|
| 1862 |
+
"loss": 2.3301,
|
| 1863 |
+
"step": 1854
|
| 1864 |
+
},
|
| 1865 |
+
{
|
| 1866 |
+
"epoch": 5.96,
|
| 1867 |
+
"learning_rate": 2.0192307692307694e-05,
|
| 1868 |
+
"loss": 2.545,
|
| 1869 |
+
"step": 1860
|
| 1870 |
+
},
|
| 1871 |
+
{
|
| 1872 |
+
"epoch": 5.98,
|
| 1873 |
+
"learning_rate": 2.0096153846153846e-05,
|
| 1874 |
+
"loss": 2.3865,
|
| 1875 |
+
"step": 1866
|
| 1876 |
+
},
|
| 1877 |
+
{
|
| 1878 |
+
"epoch": 6.0,
|
| 1879 |
+
"learning_rate": 2e-05,
|
| 1880 |
+
"loss": 2.4531,
|
| 1881 |
+
"step": 1872
|
| 1882 |
+
},
|
| 1883 |
+
{
|
| 1884 |
+
"epoch": 6.02,
|
| 1885 |
+
"learning_rate": 1.9903846153846154e-05,
|
| 1886 |
+
"loss": 2.4046,
|
| 1887 |
+
"step": 1878
|
| 1888 |
+
},
|
| 1889 |
+
{
|
| 1890 |
+
"epoch": 6.04,
|
| 1891 |
+
"learning_rate": 1.980769230769231e-05,
|
| 1892 |
+
"loss": 2.476,
|
| 1893 |
+
"step": 1884
|
| 1894 |
+
},
|
| 1895 |
+
{
|
| 1896 |
+
"epoch": 6.06,
|
| 1897 |
+
"learning_rate": 1.971153846153846e-05,
|
| 1898 |
+
"loss": 2.3547,
|
| 1899 |
+
"step": 1890
|
| 1900 |
+
},
|
| 1901 |
+
{
|
| 1902 |
+
"epoch": 6.08,
|
| 1903 |
+
"learning_rate": 1.9615384615384617e-05,
|
| 1904 |
+
"loss": 2.5396,
|
| 1905 |
+
"step": 1896
|
| 1906 |
+
},
|
| 1907 |
+
{
|
| 1908 |
+
"epoch": 6.1,
|
| 1909 |
+
"learning_rate": 1.951923076923077e-05,
|
| 1910 |
+
"loss": 2.3868,
|
| 1911 |
+
"step": 1902
|
| 1912 |
+
},
|
| 1913 |
+
{
|
| 1914 |
+
"epoch": 6.12,
|
| 1915 |
+
"learning_rate": 1.9423076923076924e-05,
|
| 1916 |
+
"loss": 2.2733,
|
| 1917 |
+
"step": 1908
|
| 1918 |
+
},
|
| 1919 |
+
{
|
| 1920 |
+
"epoch": 6.13,
|
| 1921 |
+
"learning_rate": 1.932692307692308e-05,
|
| 1922 |
+
"loss": 2.283,
|
| 1923 |
+
"step": 1914
|
| 1924 |
+
},
|
| 1925 |
+
{
|
| 1926 |
+
"epoch": 6.15,
|
| 1927 |
+
"learning_rate": 1.923076923076923e-05,
|
| 1928 |
+
"loss": 2.4231,
|
| 1929 |
+
"step": 1920
|
| 1930 |
+
},
|
| 1931 |
+
{
|
| 1932 |
+
"epoch": 6.17,
|
| 1933 |
+
"learning_rate": 1.9134615384615383e-05,
|
| 1934 |
+
"loss": 2.375,
|
| 1935 |
+
"step": 1926
|
| 1936 |
+
},
|
| 1937 |
+
{
|
| 1938 |
+
"epoch": 6.19,
|
| 1939 |
+
"learning_rate": 1.903846153846154e-05,
|
| 1940 |
+
"loss": 2.3824,
|
| 1941 |
+
"step": 1932
|
| 1942 |
+
},
|
| 1943 |
+
{
|
| 1944 |
+
"epoch": 6.21,
|
| 1945 |
+
"learning_rate": 1.894230769230769e-05,
|
| 1946 |
+
"loss": 2.3903,
|
| 1947 |
+
"step": 1938
|
| 1948 |
+
},
|
| 1949 |
+
{
|
| 1950 |
+
"epoch": 6.23,
|
| 1951 |
+
"learning_rate": 1.8846153846153846e-05,
|
| 1952 |
+
"loss": 2.2854,
|
| 1953 |
+
"step": 1944
|
| 1954 |
+
},
|
| 1955 |
+
{
|
| 1956 |
+
"epoch": 6.25,
|
| 1957 |
+
"learning_rate": 1.8750000000000002e-05,
|
| 1958 |
+
"loss": 2.2358,
|
| 1959 |
+
"step": 1950
|
| 1960 |
+
},
|
| 1961 |
+
{
|
| 1962 |
+
"epoch": 6.27,
|
| 1963 |
+
"learning_rate": 1.8653846153846154e-05,
|
| 1964 |
+
"loss": 2.4751,
|
| 1965 |
+
"step": 1956
|
| 1966 |
+
},
|
| 1967 |
+
{
|
| 1968 |
+
"epoch": 6.29,
|
| 1969 |
+
"learning_rate": 1.855769230769231e-05,
|
| 1970 |
+
"loss": 2.3434,
|
| 1971 |
+
"step": 1962
|
| 1972 |
+
},
|
| 1973 |
+
{
|
| 1974 |
+
"epoch": 6.31,
|
| 1975 |
+
"learning_rate": 1.8461538461538465e-05,
|
| 1976 |
+
"loss": 2.4088,
|
| 1977 |
+
"step": 1968
|
| 1978 |
+
},
|
| 1979 |
+
{
|
| 1980 |
+
"epoch": 6.33,
|
| 1981 |
+
"learning_rate": 1.8365384615384617e-05,
|
| 1982 |
+
"loss": 2.2746,
|
| 1983 |
+
"step": 1974
|
| 1984 |
+
},
|
| 1985 |
+
{
|
| 1986 |
+
"epoch": 6.35,
|
| 1987 |
+
"learning_rate": 1.826923076923077e-05,
|
| 1988 |
+
"loss": 2.4396,
|
| 1989 |
+
"step": 1980
|
| 1990 |
+
},
|
| 1991 |
+
{
|
| 1992 |
+
"epoch": 6.37,
|
| 1993 |
+
"learning_rate": 1.8173076923076924e-05,
|
| 1994 |
+
"loss": 2.3663,
|
| 1995 |
+
"step": 1986
|
| 1996 |
+
},
|
| 1997 |
+
{
|
| 1998 |
+
"epoch": 6.38,
|
| 1999 |
+
"learning_rate": 1.8076923076923076e-05,
|
| 2000 |
+
"loss": 2.4488,
|
| 2001 |
+
"step": 1992
|
| 2002 |
+
},
|
| 2003 |
+
{
|
| 2004 |
+
"epoch": 6.4,
|
| 2005 |
+
"learning_rate": 1.798076923076923e-05,
|
| 2006 |
+
"loss": 2.3495,
|
| 2007 |
+
"step": 1998
|
| 2008 |
+
},
|
| 2009 |
+
{
|
| 2010 |
+
"epoch": 6.42,
|
| 2011 |
+
"learning_rate": 1.7884615384615387e-05,
|
| 2012 |
+
"loss": 2.4099,
|
| 2013 |
+
"step": 2004
|
| 2014 |
+
},
|
| 2015 |
+
{
|
| 2016 |
+
"epoch": 6.44,
|
| 2017 |
+
"learning_rate": 1.778846153846154e-05,
|
| 2018 |
+
"loss": 2.4343,
|
| 2019 |
+
"step": 2010
|
| 2020 |
+
},
|
| 2021 |
+
{
|
| 2022 |
+
"epoch": 6.46,
|
| 2023 |
+
"learning_rate": 1.7692307692307694e-05,
|
| 2024 |
+
"loss": 2.3508,
|
| 2025 |
+
"step": 2016
|
| 2026 |
+
},
|
| 2027 |
+
{
|
| 2028 |
+
"epoch": 6.48,
|
| 2029 |
+
"learning_rate": 1.759615384615385e-05,
|
| 2030 |
+
"loss": 2.4724,
|
| 2031 |
+
"step": 2022
|
| 2032 |
+
},
|
| 2033 |
+
{
|
| 2034 |
+
"epoch": 6.5,
|
| 2035 |
+
"learning_rate": 1.75e-05,
|
| 2036 |
+
"loss": 2.4746,
|
| 2037 |
+
"step": 2028
|
| 2038 |
+
},
|
| 2039 |
+
{
|
| 2040 |
+
"epoch": 6.52,
|
| 2041 |
+
"learning_rate": 1.7403846153846154e-05,
|
| 2042 |
+
"loss": 2.3899,
|
| 2043 |
+
"step": 2034
|
| 2044 |
+
},
|
| 2045 |
+
{
|
| 2046 |
+
"epoch": 6.54,
|
| 2047 |
+
"learning_rate": 1.730769230769231e-05,
|
| 2048 |
+
"loss": 2.3465,
|
| 2049 |
+
"step": 2040
|
| 2050 |
+
},
|
| 2051 |
+
{
|
| 2052 |
+
"epoch": 6.56,
|
| 2053 |
+
"learning_rate": 1.721153846153846e-05,
|
| 2054 |
+
"loss": 2.3984,
|
| 2055 |
+
"step": 2046
|
| 2056 |
+
},
|
| 2057 |
+
{
|
| 2058 |
+
"epoch": 6.58,
|
| 2059 |
+
"learning_rate": 1.7115384615384617e-05,
|
| 2060 |
+
"loss": 2.3041,
|
| 2061 |
+
"step": 2052
|
| 2062 |
+
},
|
| 2063 |
+
{
|
| 2064 |
+
"epoch": 6.6,
|
| 2065 |
+
"learning_rate": 1.7019230769230772e-05,
|
| 2066 |
+
"loss": 2.3035,
|
| 2067 |
+
"step": 2058
|
| 2068 |
+
},
|
| 2069 |
+
{
|
| 2070 |
+
"epoch": 6.62,
|
| 2071 |
+
"learning_rate": 1.6923076923076924e-05,
|
| 2072 |
+
"loss": 2.2846,
|
| 2073 |
+
"step": 2064
|
| 2074 |
+
},
|
| 2075 |
+
{
|
| 2076 |
+
"epoch": 6.63,
|
| 2077 |
+
"learning_rate": 1.682692307692308e-05,
|
| 2078 |
+
"loss": 2.4577,
|
| 2079 |
+
"step": 2070
|
| 2080 |
+
},
|
| 2081 |
+
{
|
| 2082 |
+
"epoch": 6.65,
|
| 2083 |
+
"learning_rate": 1.673076923076923e-05,
|
| 2084 |
+
"loss": 2.3338,
|
| 2085 |
+
"step": 2076
|
| 2086 |
+
},
|
| 2087 |
+
{
|
| 2088 |
+
"epoch": 6.67,
|
| 2089 |
+
"learning_rate": 1.6634615384615384e-05,
|
| 2090 |
+
"loss": 2.3623,
|
| 2091 |
+
"step": 2082
|
| 2092 |
+
},
|
| 2093 |
+
{
|
| 2094 |
+
"epoch": 6.69,
|
| 2095 |
+
"learning_rate": 1.653846153846154e-05,
|
| 2096 |
+
"loss": 2.3601,
|
| 2097 |
+
"step": 2088
|
| 2098 |
+
},
|
| 2099 |
+
{
|
| 2100 |
+
"epoch": 6.71,
|
| 2101 |
+
"learning_rate": 1.6442307692307695e-05,
|
| 2102 |
+
"loss": 2.34,
|
| 2103 |
+
"step": 2094
|
| 2104 |
+
},
|
| 2105 |
+
{
|
| 2106 |
+
"epoch": 6.73,
|
| 2107 |
+
"learning_rate": 1.6346153846153847e-05,
|
| 2108 |
+
"loss": 2.3451,
|
| 2109 |
+
"step": 2100
|
| 2110 |
+
},
|
| 2111 |
+
{
|
| 2112 |
+
"epoch": 6.75,
|
| 2113 |
+
"learning_rate": 1.6250000000000002e-05,
|
| 2114 |
+
"loss": 2.3392,
|
| 2115 |
+
"step": 2106
|
| 2116 |
+
},
|
| 2117 |
+
{
|
| 2118 |
+
"epoch": 6.77,
|
| 2119 |
+
"learning_rate": 1.6153846153846154e-05,
|
| 2120 |
+
"loss": 2.4011,
|
| 2121 |
+
"step": 2112
|
| 2122 |
+
},
|
| 2123 |
+
{
|
| 2124 |
+
"epoch": 6.79,
|
| 2125 |
+
"learning_rate": 1.605769230769231e-05,
|
| 2126 |
+
"loss": 2.382,
|
| 2127 |
+
"step": 2118
|
| 2128 |
+
},
|
| 2129 |
+
{
|
| 2130 |
+
"epoch": 6.81,
|
| 2131 |
+
"learning_rate": 1.596153846153846e-05,
|
| 2132 |
+
"loss": 2.4005,
|
| 2133 |
+
"step": 2124
|
| 2134 |
+
},
|
| 2135 |
+
{
|
| 2136 |
+
"epoch": 6.83,
|
| 2137 |
+
"learning_rate": 1.5865384615384617e-05,
|
| 2138 |
+
"loss": 2.3398,
|
| 2139 |
+
"step": 2130
|
| 2140 |
+
},
|
| 2141 |
+
{
|
| 2142 |
+
"epoch": 6.85,
|
| 2143 |
+
"learning_rate": 1.576923076923077e-05,
|
| 2144 |
+
"loss": 2.3156,
|
| 2145 |
+
"step": 2136
|
| 2146 |
+
},
|
| 2147 |
+
{
|
| 2148 |
+
"epoch": 6.87,
|
| 2149 |
+
"learning_rate": 1.5673076923076924e-05,
|
| 2150 |
+
"loss": 2.3249,
|
| 2151 |
+
"step": 2142
|
| 2152 |
+
},
|
| 2153 |
+
{
|
| 2154 |
+
"epoch": 6.88,
|
| 2155 |
+
"learning_rate": 1.5576923076923076e-05,
|
| 2156 |
+
"loss": 2.2891,
|
| 2157 |
+
"step": 2148
|
| 2158 |
+
},
|
| 2159 |
+
{
|
| 2160 |
+
"epoch": 6.9,
|
| 2161 |
+
"learning_rate": 1.5480769230769232e-05,
|
| 2162 |
+
"loss": 2.3154,
|
| 2163 |
+
"step": 2154
|
| 2164 |
+
},
|
| 2165 |
+
{
|
| 2166 |
+
"epoch": 6.92,
|
| 2167 |
+
"learning_rate": 1.5384615384615387e-05,
|
| 2168 |
+
"loss": 2.3642,
|
| 2169 |
+
"step": 2160
|
| 2170 |
+
},
|
| 2171 |
+
{
|
| 2172 |
+
"epoch": 6.94,
|
| 2173 |
+
"learning_rate": 1.528846153846154e-05,
|
| 2174 |
+
"loss": 2.3303,
|
| 2175 |
+
"step": 2166
|
| 2176 |
+
},
|
| 2177 |
+
{
|
| 2178 |
+
"epoch": 6.96,
|
| 2179 |
+
"learning_rate": 1.5192307692307691e-05,
|
| 2180 |
+
"loss": 2.4364,
|
| 2181 |
+
"step": 2172
|
| 2182 |
+
},
|
| 2183 |
+
{
|
| 2184 |
+
"epoch": 6.98,
|
| 2185 |
+
"learning_rate": 1.5096153846153847e-05,
|
| 2186 |
+
"loss": 2.4258,
|
| 2187 |
+
"step": 2178
|
| 2188 |
+
},
|
| 2189 |
+
{
|
| 2190 |
+
"epoch": 7.0,
|
| 2191 |
+
"learning_rate": 1.5e-05,
|
| 2192 |
+
"loss": 2.3566,
|
| 2193 |
+
"step": 2184
|
| 2194 |
+
},
|
| 2195 |
+
{
|
| 2196 |
+
"epoch": 7.02,
|
| 2197 |
+
"learning_rate": 1.4903846153846154e-05,
|
| 2198 |
+
"loss": 2.2916,
|
| 2199 |
+
"step": 2190
|
| 2200 |
+
},
|
| 2201 |
+
{
|
| 2202 |
+
"epoch": 7.04,
|
| 2203 |
+
"learning_rate": 1.4807692307692308e-05,
|
| 2204 |
+
"loss": 2.3829,
|
| 2205 |
+
"step": 2196
|
| 2206 |
+
},
|
| 2207 |
+
{
|
| 2208 |
+
"epoch": 7.06,
|
| 2209 |
+
"learning_rate": 1.4711538461538463e-05,
|
| 2210 |
+
"loss": 2.242,
|
| 2211 |
+
"step": 2202
|
| 2212 |
+
},
|
| 2213 |
+
{
|
| 2214 |
+
"epoch": 7.08,
|
| 2215 |
+
"learning_rate": 1.4615384615384617e-05,
|
| 2216 |
+
"loss": 2.2546,
|
| 2217 |
+
"step": 2208
|
| 2218 |
+
},
|
| 2219 |
+
{
|
| 2220 |
+
"epoch": 7.1,
|
| 2221 |
+
"learning_rate": 1.451923076923077e-05,
|
| 2222 |
+
"loss": 2.1526,
|
| 2223 |
+
"step": 2214
|
| 2224 |
+
},
|
| 2225 |
+
{
|
| 2226 |
+
"epoch": 7.12,
|
| 2227 |
+
"learning_rate": 1.4423076923076923e-05,
|
| 2228 |
+
"loss": 2.2073,
|
| 2229 |
+
"step": 2220
|
| 2230 |
+
},
|
| 2231 |
+
{
|
| 2232 |
+
"epoch": 7.13,
|
| 2233 |
+
"learning_rate": 1.4326923076923076e-05,
|
| 2234 |
+
"loss": 2.2992,
|
| 2235 |
+
"step": 2226
|
| 2236 |
+
},
|
| 2237 |
+
{
|
| 2238 |
+
"epoch": 7.15,
|
| 2239 |
+
"learning_rate": 1.423076923076923e-05,
|
| 2240 |
+
"loss": 2.2544,
|
| 2241 |
+
"step": 2232
|
| 2242 |
+
},
|
| 2243 |
+
{
|
| 2244 |
+
"epoch": 7.17,
|
| 2245 |
+
"learning_rate": 1.4134615384615386e-05,
|
| 2246 |
+
"loss": 2.2632,
|
| 2247 |
+
"step": 2238
|
| 2248 |
+
},
|
| 2249 |
+
{
|
| 2250 |
+
"epoch": 7.19,
|
| 2251 |
+
"learning_rate": 1.403846153846154e-05,
|
| 2252 |
+
"loss": 2.3158,
|
| 2253 |
+
"step": 2244
|
| 2254 |
+
},
|
| 2255 |
+
{
|
| 2256 |
+
"epoch": 7.21,
|
| 2257 |
+
"learning_rate": 1.3942307692307693e-05,
|
| 2258 |
+
"loss": 2.2981,
|
| 2259 |
+
"step": 2250
|
| 2260 |
+
},
|
| 2261 |
+
{
|
| 2262 |
+
"epoch": 7.23,
|
| 2263 |
+
"learning_rate": 1.3846153846153847e-05,
|
| 2264 |
+
"loss": 2.2468,
|
| 2265 |
+
"step": 2256
|
| 2266 |
+
},
|
| 2267 |
+
{
|
| 2268 |
+
"epoch": 7.25,
|
| 2269 |
+
"learning_rate": 1.3750000000000002e-05,
|
| 2270 |
+
"loss": 2.4333,
|
| 2271 |
+
"step": 2262
|
| 2272 |
+
},
|
| 2273 |
+
{
|
| 2274 |
+
"epoch": 7.27,
|
| 2275 |
+
"learning_rate": 1.3653846153846153e-05,
|
| 2276 |
+
"loss": 2.3005,
|
| 2277 |
+
"step": 2268
|
| 2278 |
+
},
|
| 2279 |
+
{
|
| 2280 |
+
"epoch": 7.29,
|
| 2281 |
+
"learning_rate": 1.3557692307692308e-05,
|
| 2282 |
+
"loss": 2.2629,
|
| 2283 |
+
"step": 2274
|
| 2284 |
+
},
|
| 2285 |
+
{
|
| 2286 |
+
"epoch": 7.31,
|
| 2287 |
+
"learning_rate": 1.3461538461538462e-05,
|
| 2288 |
+
"loss": 2.3122,
|
| 2289 |
+
"step": 2280
|
| 2290 |
+
},
|
| 2291 |
+
{
|
| 2292 |
+
"epoch": 7.33,
|
| 2293 |
+
"learning_rate": 1.3365384615384615e-05,
|
| 2294 |
+
"loss": 2.3257,
|
| 2295 |
+
"step": 2286
|
| 2296 |
+
},
|
| 2297 |
+
{
|
| 2298 |
+
"epoch": 7.35,
|
| 2299 |
+
"learning_rate": 1.3269230769230769e-05,
|
| 2300 |
+
"loss": 2.4191,
|
| 2301 |
+
"step": 2292
|
| 2302 |
+
},
|
| 2303 |
+
{
|
| 2304 |
+
"epoch": 7.37,
|
| 2305 |
+
"learning_rate": 1.3173076923076925e-05,
|
| 2306 |
+
"loss": 2.2431,
|
| 2307 |
+
"step": 2298
|
| 2308 |
+
},
|
| 2309 |
+
{
|
| 2310 |
+
"epoch": 7.38,
|
| 2311 |
+
"learning_rate": 1.3076923076923078e-05,
|
| 2312 |
+
"loss": 2.3716,
|
| 2313 |
+
"step": 2304
|
| 2314 |
+
},
|
| 2315 |
+
{
|
| 2316 |
+
"epoch": 7.4,
|
| 2317 |
+
"learning_rate": 1.2980769230769232e-05,
|
| 2318 |
+
"loss": 2.1685,
|
| 2319 |
+
"step": 2310
|
| 2320 |
+
},
|
| 2321 |
+
{
|
| 2322 |
+
"epoch": 7.42,
|
| 2323 |
+
"learning_rate": 1.2884615384615384e-05,
|
| 2324 |
+
"loss": 2.2519,
|
| 2325 |
+
"step": 2316
|
| 2326 |
+
},
|
| 2327 |
+
{
|
| 2328 |
+
"epoch": 7.44,
|
| 2329 |
+
"learning_rate": 1.2788461538461538e-05,
|
| 2330 |
+
"loss": 2.2242,
|
| 2331 |
+
"step": 2322
|
| 2332 |
+
},
|
| 2333 |
+
{
|
| 2334 |
+
"epoch": 7.46,
|
| 2335 |
+
"learning_rate": 1.2692307692307691e-05,
|
| 2336 |
+
"loss": 2.3481,
|
| 2337 |
+
"step": 2328
|
| 2338 |
+
},
|
| 2339 |
+
{
|
| 2340 |
+
"epoch": 7.48,
|
| 2341 |
+
"learning_rate": 1.2596153846153847e-05,
|
| 2342 |
+
"loss": 2.3117,
|
| 2343 |
+
"step": 2334
|
| 2344 |
+
},
|
| 2345 |
+
{
|
| 2346 |
+
"epoch": 7.5,
|
| 2347 |
+
"learning_rate": 1.25e-05,
|
| 2348 |
+
"loss": 2.2924,
|
| 2349 |
+
"step": 2340
|
| 2350 |
+
},
|
| 2351 |
+
{
|
| 2352 |
+
"epoch": 7.52,
|
| 2353 |
+
"learning_rate": 1.2403846153846154e-05,
|
| 2354 |
+
"loss": 2.2933,
|
| 2355 |
+
"step": 2346
|
| 2356 |
+
},
|
| 2357 |
+
{
|
| 2358 |
+
"epoch": 7.54,
|
| 2359 |
+
"learning_rate": 1.230769230769231e-05,
|
| 2360 |
+
"loss": 2.298,
|
| 2361 |
+
"step": 2352
|
| 2362 |
+
},
|
| 2363 |
+
{
|
| 2364 |
+
"epoch": 7.56,
|
| 2365 |
+
"learning_rate": 1.2211538461538462e-05,
|
| 2366 |
+
"loss": 2.2145,
|
| 2367 |
+
"step": 2358
|
| 2368 |
+
},
|
| 2369 |
+
{
|
| 2370 |
+
"epoch": 7.58,
|
| 2371 |
+
"learning_rate": 1.2115384615384615e-05,
|
| 2372 |
+
"loss": 2.3118,
|
| 2373 |
+
"step": 2364
|
| 2374 |
+
},
|
| 2375 |
+
{
|
| 2376 |
+
"epoch": 7.6,
|
| 2377 |
+
"learning_rate": 1.2019230769230771e-05,
|
| 2378 |
+
"loss": 2.3132,
|
| 2379 |
+
"step": 2370
|
| 2380 |
+
},
|
| 2381 |
+
{
|
| 2382 |
+
"epoch": 7.62,
|
| 2383 |
+
"learning_rate": 1.1923076923076925e-05,
|
| 2384 |
+
"loss": 2.4653,
|
| 2385 |
+
"step": 2376
|
| 2386 |
+
},
|
| 2387 |
+
{
|
| 2388 |
+
"epoch": 7.63,
|
| 2389 |
+
"learning_rate": 1.1826923076923077e-05,
|
| 2390 |
+
"loss": 2.2083,
|
| 2391 |
+
"step": 2382
|
| 2392 |
+
},
|
| 2393 |
+
{
|
| 2394 |
+
"epoch": 7.65,
|
| 2395 |
+
"learning_rate": 1.1730769230769232e-05,
|
| 2396 |
+
"loss": 2.2876,
|
| 2397 |
+
"step": 2388
|
| 2398 |
+
},
|
| 2399 |
+
{
|
| 2400 |
+
"epoch": 7.67,
|
| 2401 |
+
"learning_rate": 1.1634615384615386e-05,
|
| 2402 |
+
"loss": 2.2115,
|
| 2403 |
+
"step": 2394
|
| 2404 |
+
},
|
| 2405 |
+
{
|
| 2406 |
+
"epoch": 7.69,
|
| 2407 |
+
"learning_rate": 1.153846153846154e-05,
|
| 2408 |
+
"loss": 2.3116,
|
| 2409 |
+
"step": 2400
|
| 2410 |
+
},
|
| 2411 |
+
{
|
| 2412 |
+
"epoch": 7.71,
|
| 2413 |
+
"learning_rate": 1.1442307692307693e-05,
|
| 2414 |
+
"loss": 2.2858,
|
| 2415 |
+
"step": 2406
|
| 2416 |
+
},
|
| 2417 |
+
{
|
| 2418 |
+
"epoch": 7.73,
|
| 2419 |
+
"learning_rate": 1.1346153846153847e-05,
|
| 2420 |
+
"loss": 2.2525,
|
| 2421 |
+
"step": 2412
|
| 2422 |
+
},
|
| 2423 |
+
{
|
| 2424 |
+
"epoch": 7.75,
|
| 2425 |
+
"learning_rate": 1.125e-05,
|
| 2426 |
+
"loss": 2.3215,
|
| 2427 |
+
"step": 2418
|
| 2428 |
+
},
|
| 2429 |
+
{
|
| 2430 |
+
"epoch": 7.77,
|
| 2431 |
+
"learning_rate": 1.1153846153846154e-05,
|
| 2432 |
+
"loss": 2.3884,
|
| 2433 |
+
"step": 2424
|
| 2434 |
+
},
|
| 2435 |
+
{
|
| 2436 |
+
"epoch": 7.79,
|
| 2437 |
+
"learning_rate": 1.1057692307692308e-05,
|
| 2438 |
+
"loss": 2.3767,
|
| 2439 |
+
"step": 2430
|
| 2440 |
+
},
|
| 2441 |
+
{
|
| 2442 |
+
"epoch": 7.81,
|
| 2443 |
+
"learning_rate": 1.0961538461538462e-05,
|
| 2444 |
+
"loss": 2.1214,
|
| 2445 |
+
"step": 2436
|
| 2446 |
+
},
|
| 2447 |
+
{
|
| 2448 |
+
"epoch": 7.83,
|
| 2449 |
+
"learning_rate": 1.0865384615384616e-05,
|
| 2450 |
+
"loss": 2.3003,
|
| 2451 |
+
"step": 2442
|
| 2452 |
+
},
|
| 2453 |
+
{
|
| 2454 |
+
"epoch": 7.85,
|
| 2455 |
+
"learning_rate": 1.0769230769230771e-05,
|
| 2456 |
+
"loss": 2.426,
|
| 2457 |
+
"step": 2448
|
| 2458 |
+
},
|
| 2459 |
+
{
|
| 2460 |
+
"epoch": 7.87,
|
| 2461 |
+
"learning_rate": 1.0673076923076923e-05,
|
| 2462 |
+
"loss": 2.2785,
|
| 2463 |
+
"step": 2454
|
| 2464 |
+
},
|
| 2465 |
+
{
|
| 2466 |
+
"epoch": 7.88,
|
| 2467 |
+
"learning_rate": 1.0576923076923077e-05,
|
| 2468 |
+
"loss": 2.3252,
|
| 2469 |
+
"step": 2460
|
| 2470 |
+
},
|
| 2471 |
+
{
|
| 2472 |
+
"epoch": 7.9,
|
| 2473 |
+
"learning_rate": 1.0480769230769232e-05,
|
| 2474 |
+
"loss": 2.2636,
|
| 2475 |
+
"step": 2466
|
| 2476 |
+
},
|
| 2477 |
+
{
|
| 2478 |
+
"epoch": 7.92,
|
| 2479 |
+
"learning_rate": 1.0384615384615386e-05,
|
| 2480 |
+
"loss": 2.2732,
|
| 2481 |
+
"step": 2472
|
| 2482 |
+
},
|
| 2483 |
+
{
|
| 2484 |
+
"epoch": 7.94,
|
| 2485 |
+
"learning_rate": 1.0288461538461538e-05,
|
| 2486 |
+
"loss": 2.1934,
|
| 2487 |
+
"step": 2478
|
| 2488 |
+
},
|
| 2489 |
+
{
|
| 2490 |
+
"epoch": 7.96,
|
| 2491 |
+
"learning_rate": 1.0192307692307693e-05,
|
| 2492 |
+
"loss": 2.2967,
|
| 2493 |
+
"step": 2484
|
| 2494 |
+
},
|
| 2495 |
+
{
|
| 2496 |
+
"epoch": 7.98,
|
| 2497 |
+
"learning_rate": 1.0096153846153847e-05,
|
| 2498 |
+
"loss": 2.2939,
|
| 2499 |
+
"step": 2490
|
| 2500 |
+
},
|
| 2501 |
+
{
|
| 2502 |
+
"epoch": 8.0,
|
| 2503 |
+
"learning_rate": 1e-05,
|
| 2504 |
+
"loss": 2.3393,
|
| 2505 |
+
"step": 2496
|
| 2506 |
+
},
|
| 2507 |
+
{
|
| 2508 |
+
"epoch": 8.02,
|
| 2509 |
+
"learning_rate": 9.903846153846155e-06,
|
| 2510 |
+
"loss": 2.1356,
|
| 2511 |
+
"step": 2502
|
| 2512 |
+
},
|
| 2513 |
+
{
|
| 2514 |
+
"epoch": 8.04,
|
| 2515 |
+
"learning_rate": 9.807692307692308e-06,
|
| 2516 |
+
"loss": 2.2868,
|
| 2517 |
+
"step": 2508
|
| 2518 |
+
},
|
| 2519 |
+
{
|
| 2520 |
+
"epoch": 8.06,
|
| 2521 |
+
"learning_rate": 9.711538461538462e-06,
|
| 2522 |
+
"loss": 2.2453,
|
| 2523 |
+
"step": 2514
|
| 2524 |
+
},
|
| 2525 |
+
{
|
| 2526 |
+
"epoch": 8.08,
|
| 2527 |
+
"learning_rate": 9.615384615384616e-06,
|
| 2528 |
+
"loss": 2.2619,
|
| 2529 |
+
"step": 2520
|
| 2530 |
+
},
|
| 2531 |
+
{
|
| 2532 |
+
"epoch": 8.1,
|
| 2533 |
+
"learning_rate": 9.51923076923077e-06,
|
| 2534 |
+
"loss": 2.2185,
|
| 2535 |
+
"step": 2526
|
| 2536 |
+
},
|
| 2537 |
+
{
|
| 2538 |
+
"epoch": 8.12,
|
| 2539 |
+
"learning_rate": 9.423076923076923e-06,
|
| 2540 |
+
"loss": 2.2548,
|
| 2541 |
+
"step": 2532
|
| 2542 |
+
},
|
| 2543 |
+
{
|
| 2544 |
+
"epoch": 8.13,
|
| 2545 |
+
"learning_rate": 9.326923076923077e-06,
|
| 2546 |
+
"loss": 2.2613,
|
| 2547 |
+
"step": 2538
|
| 2548 |
+
},
|
| 2549 |
+
{
|
| 2550 |
+
"epoch": 8.15,
|
| 2551 |
+
"learning_rate": 9.230769230769232e-06,
|
| 2552 |
+
"loss": 2.3997,
|
| 2553 |
+
"step": 2544
|
| 2554 |
+
},
|
| 2555 |
+
{
|
| 2556 |
+
"epoch": 8.17,
|
| 2557 |
+
"learning_rate": 9.134615384615384e-06,
|
| 2558 |
+
"loss": 2.1023,
|
| 2559 |
+
"step": 2550
|
| 2560 |
+
},
|
| 2561 |
+
{
|
| 2562 |
+
"epoch": 8.19,
|
| 2563 |
+
"learning_rate": 9.038461538461538e-06,
|
| 2564 |
+
"loss": 2.2411,
|
| 2565 |
+
"step": 2556
|
| 2566 |
+
},
|
| 2567 |
+
{
|
| 2568 |
+
"epoch": 8.21,
|
| 2569 |
+
"learning_rate": 8.942307692307693e-06,
|
| 2570 |
+
"loss": 2.1498,
|
| 2571 |
+
"step": 2562
|
| 2572 |
+
},
|
| 2573 |
+
{
|
| 2574 |
+
"epoch": 8.23,
|
| 2575 |
+
"learning_rate": 8.846153846153847e-06,
|
| 2576 |
+
"loss": 2.2051,
|
| 2577 |
+
"step": 2568
|
| 2578 |
+
},
|
| 2579 |
+
{
|
| 2580 |
+
"epoch": 8.25,
|
| 2581 |
+
"learning_rate": 8.75e-06,
|
| 2582 |
+
"loss": 2.1779,
|
| 2583 |
+
"step": 2574
|
| 2584 |
+
},
|
| 2585 |
+
{
|
| 2586 |
+
"epoch": 8.27,
|
| 2587 |
+
"learning_rate": 8.653846153846155e-06,
|
| 2588 |
+
"loss": 2.2369,
|
| 2589 |
+
"step": 2580
|
| 2590 |
+
},
|
| 2591 |
+
{
|
| 2592 |
+
"epoch": 8.29,
|
| 2593 |
+
"learning_rate": 8.557692307692308e-06,
|
| 2594 |
+
"loss": 2.1945,
|
| 2595 |
+
"step": 2586
|
| 2596 |
+
},
|
| 2597 |
+
{
|
| 2598 |
+
"epoch": 8.31,
|
| 2599 |
+
"learning_rate": 8.461538461538462e-06,
|
| 2600 |
+
"loss": 2.1002,
|
| 2601 |
+
"step": 2592
|
| 2602 |
+
},
|
| 2603 |
+
{
|
| 2604 |
+
"epoch": 8.33,
|
| 2605 |
+
"learning_rate": 8.365384615384616e-06,
|
| 2606 |
+
"loss": 2.1385,
|
| 2607 |
+
"step": 2598
|
| 2608 |
+
},
|
| 2609 |
+
{
|
| 2610 |
+
"epoch": 8.35,
|
| 2611 |
+
"learning_rate": 8.26923076923077e-06,
|
| 2612 |
+
"loss": 2.3863,
|
| 2613 |
+
"step": 2604
|
| 2614 |
+
},
|
| 2615 |
+
{
|
| 2616 |
+
"epoch": 8.37,
|
| 2617 |
+
"learning_rate": 8.173076923076923e-06,
|
| 2618 |
+
"loss": 2.2337,
|
| 2619 |
+
"step": 2610
|
| 2620 |
+
},
|
| 2621 |
+
{
|
| 2622 |
+
"epoch": 8.38,
|
| 2623 |
+
"learning_rate": 8.076923076923077e-06,
|
| 2624 |
+
"loss": 2.2459,
|
| 2625 |
+
"step": 2616
|
| 2626 |
+
},
|
| 2627 |
+
{
|
| 2628 |
+
"epoch": 8.4,
|
| 2629 |
+
"learning_rate": 7.98076923076923e-06,
|
| 2630 |
+
"loss": 2.1587,
|
| 2631 |
+
"step": 2622
|
| 2632 |
+
},
|
| 2633 |
+
{
|
| 2634 |
+
"epoch": 8.42,
|
| 2635 |
+
"learning_rate": 7.884615384615384e-06,
|
| 2636 |
+
"loss": 2.2672,
|
| 2637 |
+
"step": 2628
|
| 2638 |
+
},
|
| 2639 |
+
{
|
| 2640 |
+
"epoch": 8.44,
|
| 2641 |
+
"learning_rate": 7.788461538461538e-06,
|
| 2642 |
+
"loss": 2.314,
|
| 2643 |
+
"step": 2634
|
| 2644 |
+
},
|
| 2645 |
+
{
|
| 2646 |
+
"epoch": 8.46,
|
| 2647 |
+
"learning_rate": 7.692307692307694e-06,
|
| 2648 |
+
"loss": 2.2892,
|
| 2649 |
+
"step": 2640
|
| 2650 |
+
},
|
| 2651 |
+
{
|
| 2652 |
+
"epoch": 8.48,
|
| 2653 |
+
"learning_rate": 7.596153846153846e-06,
|
| 2654 |
+
"loss": 2.2787,
|
| 2655 |
+
"step": 2646
|
| 2656 |
+
},
|
| 2657 |
+
{
|
| 2658 |
+
"epoch": 8.5,
|
| 2659 |
+
"learning_rate": 7.5e-06,
|
| 2660 |
+
"loss": 2.3972,
|
| 2661 |
+
"step": 2652
|
| 2662 |
+
},
|
| 2663 |
+
{
|
| 2664 |
+
"epoch": 8.52,
|
| 2665 |
+
"learning_rate": 7.403846153846154e-06,
|
| 2666 |
+
"loss": 2.3257,
|
| 2667 |
+
"step": 2658
|
| 2668 |
+
},
|
| 2669 |
+
{
|
| 2670 |
+
"epoch": 8.54,
|
| 2671 |
+
"learning_rate": 7.3076923076923085e-06,
|
| 2672 |
+
"loss": 2.2807,
|
| 2673 |
+
"step": 2664
|
| 2674 |
+
},
|
| 2675 |
+
{
|
| 2676 |
+
"epoch": 8.56,
|
| 2677 |
+
"learning_rate": 7.211538461538461e-06,
|
| 2678 |
+
"loss": 2.2321,
|
| 2679 |
+
"step": 2670
|
| 2680 |
+
},
|
| 2681 |
+
{
|
| 2682 |
+
"epoch": 8.58,
|
| 2683 |
+
"learning_rate": 7.115384615384615e-06,
|
| 2684 |
+
"loss": 2.1732,
|
| 2685 |
+
"step": 2676
|
| 2686 |
+
},
|
| 2687 |
+
{
|
| 2688 |
+
"epoch": 8.6,
|
| 2689 |
+
"learning_rate": 7.01923076923077e-06,
|
| 2690 |
+
"loss": 2.2577,
|
| 2691 |
+
"step": 2682
|
| 2692 |
+
},
|
| 2693 |
+
{
|
| 2694 |
+
"epoch": 8.62,
|
| 2695 |
+
"learning_rate": 6.923076923076923e-06,
|
| 2696 |
+
"loss": 2.2655,
|
| 2697 |
+
"step": 2688
|
| 2698 |
+
},
|
| 2699 |
+
{
|
| 2700 |
+
"epoch": 8.63,
|
| 2701 |
+
"learning_rate": 6.826923076923076e-06,
|
| 2702 |
+
"loss": 2.2591,
|
| 2703 |
+
"step": 2694
|
| 2704 |
+
},
|
| 2705 |
+
{
|
| 2706 |
+
"epoch": 8.65,
|
| 2707 |
+
"learning_rate": 6.730769230769231e-06,
|
| 2708 |
+
"loss": 2.1755,
|
| 2709 |
+
"step": 2700
|
| 2710 |
+
},
|
| 2711 |
+
{
|
| 2712 |
+
"epoch": 8.67,
|
| 2713 |
+
"learning_rate": 6.6346153846153846e-06,
|
| 2714 |
+
"loss": 2.286,
|
| 2715 |
+
"step": 2706
|
| 2716 |
+
},
|
| 2717 |
+
{
|
| 2718 |
+
"epoch": 8.69,
|
| 2719 |
+
"learning_rate": 6.538461538461539e-06,
|
| 2720 |
+
"loss": 2.1993,
|
| 2721 |
+
"step": 2712
|
| 2722 |
+
},
|
| 2723 |
+
{
|
| 2724 |
+
"epoch": 8.71,
|
| 2725 |
+
"learning_rate": 6.442307692307692e-06,
|
| 2726 |
+
"loss": 2.416,
|
| 2727 |
+
"step": 2718
|
| 2728 |
+
},
|
| 2729 |
+
{
|
| 2730 |
+
"epoch": 8.73,
|
| 2731 |
+
"learning_rate": 6.346153846153846e-06,
|
| 2732 |
+
"loss": 2.1154,
|
| 2733 |
+
"step": 2724
|
| 2734 |
+
},
|
| 2735 |
+
{
|
| 2736 |
+
"epoch": 8.75,
|
| 2737 |
+
"learning_rate": 6.25e-06,
|
| 2738 |
+
"loss": 2.102,
|
| 2739 |
+
"step": 2730
|
| 2740 |
+
},
|
| 2741 |
+
{
|
| 2742 |
+
"epoch": 8.77,
|
| 2743 |
+
"learning_rate": 6.153846153846155e-06,
|
| 2744 |
+
"loss": 2.2151,
|
| 2745 |
+
"step": 2736
|
| 2746 |
+
},
|
| 2747 |
+
{
|
| 2748 |
+
"epoch": 8.79,
|
| 2749 |
+
"learning_rate": 6.057692307692308e-06,
|
| 2750 |
+
"loss": 2.2409,
|
| 2751 |
+
"step": 2742
|
| 2752 |
+
},
|
| 2753 |
+
{
|
| 2754 |
+
"epoch": 8.81,
|
| 2755 |
+
"learning_rate": 5.961538461538462e-06,
|
| 2756 |
+
"loss": 2.2415,
|
| 2757 |
+
"step": 2748
|
| 2758 |
+
},
|
| 2759 |
+
{
|
| 2760 |
+
"epoch": 8.83,
|
| 2761 |
+
"learning_rate": 5.865384615384616e-06,
|
| 2762 |
+
"loss": 2.3179,
|
| 2763 |
+
"step": 2754
|
| 2764 |
+
},
|
| 2765 |
+
{
|
| 2766 |
+
"epoch": 8.85,
|
| 2767 |
+
"learning_rate": 5.76923076923077e-06,
|
| 2768 |
+
"loss": 2.1671,
|
| 2769 |
+
"step": 2760
|
| 2770 |
+
},
|
| 2771 |
+
{
|
| 2772 |
+
"epoch": 8.87,
|
| 2773 |
+
"learning_rate": 5.6730769230769235e-06,
|
| 2774 |
+
"loss": 2.2679,
|
| 2775 |
+
"step": 2766
|
| 2776 |
+
},
|
| 2777 |
+
{
|
| 2778 |
+
"epoch": 8.88,
|
| 2779 |
+
"learning_rate": 5.576923076923077e-06,
|
| 2780 |
+
"loss": 2.0876,
|
| 2781 |
+
"step": 2772
|
| 2782 |
+
},
|
| 2783 |
+
{
|
| 2784 |
+
"epoch": 8.9,
|
| 2785 |
+
"learning_rate": 5.480769230769231e-06,
|
| 2786 |
+
"loss": 2.2468,
|
| 2787 |
+
"step": 2778
|
| 2788 |
+
},
|
| 2789 |
+
{
|
| 2790 |
+
"epoch": 8.92,
|
| 2791 |
+
"learning_rate": 5.3846153846153855e-06,
|
| 2792 |
+
"loss": 2.0652,
|
| 2793 |
+
"step": 2784
|
| 2794 |
+
},
|
| 2795 |
+
{
|
| 2796 |
+
"epoch": 8.94,
|
| 2797 |
+
"learning_rate": 5.288461538461538e-06,
|
| 2798 |
+
"loss": 2.3457,
|
| 2799 |
+
"step": 2790
|
| 2800 |
+
},
|
| 2801 |
+
{
|
| 2802 |
+
"epoch": 8.96,
|
| 2803 |
+
"learning_rate": 5.192307692307693e-06,
|
| 2804 |
+
"loss": 2.165,
|
| 2805 |
+
"step": 2796
|
| 2806 |
+
},
|
| 2807 |
+
{
|
| 2808 |
+
"epoch": 8.98,
|
| 2809 |
+
"learning_rate": 5.096153846153847e-06,
|
| 2810 |
+
"loss": 2.3123,
|
| 2811 |
+
"step": 2802
|
| 2812 |
+
},
|
| 2813 |
+
{
|
| 2814 |
+
"epoch": 9.0,
|
| 2815 |
+
"learning_rate": 5e-06,
|
| 2816 |
+
"loss": 2.1797,
|
| 2817 |
+
"step": 2808
|
| 2818 |
+
},
|
| 2819 |
+
{
|
| 2820 |
+
"epoch": 9.02,
|
| 2821 |
+
"learning_rate": 4.903846153846154e-06,
|
| 2822 |
+
"loss": 2.1129,
|
| 2823 |
+
"step": 2814
|
| 2824 |
+
},
|
| 2825 |
+
{
|
| 2826 |
+
"epoch": 9.04,
|
| 2827 |
+
"learning_rate": 4.807692307692308e-06,
|
| 2828 |
+
"loss": 2.1505,
|
| 2829 |
+
"step": 2820
|
| 2830 |
+
},
|
| 2831 |
+
{
|
| 2832 |
+
"epoch": 9.06,
|
| 2833 |
+
"learning_rate": 4.711538461538462e-06,
|
| 2834 |
+
"loss": 2.0979,
|
| 2835 |
+
"step": 2826
|
| 2836 |
+
},
|
| 2837 |
+
{
|
| 2838 |
+
"epoch": 9.08,
|
| 2839 |
+
"learning_rate": 4.615384615384616e-06,
|
| 2840 |
+
"loss": 2.1762,
|
| 2841 |
+
"step": 2832
|
| 2842 |
+
},
|
| 2843 |
+
{
|
| 2844 |
+
"epoch": 9.1,
|
| 2845 |
+
"learning_rate": 4.519230769230769e-06,
|
| 2846 |
+
"loss": 2.2454,
|
| 2847 |
+
"step": 2838
|
| 2848 |
+
},
|
| 2849 |
+
{
|
| 2850 |
+
"epoch": 9.12,
|
| 2851 |
+
"learning_rate": 4.423076923076924e-06,
|
| 2852 |
+
"loss": 2.0327,
|
| 2853 |
+
"step": 2844
|
| 2854 |
+
},
|
| 2855 |
+
{
|
| 2856 |
+
"epoch": 9.13,
|
| 2857 |
+
"learning_rate": 4.326923076923077e-06,
|
| 2858 |
+
"loss": 2.2497,
|
| 2859 |
+
"step": 2850
|
| 2860 |
+
},
|
| 2861 |
+
{
|
| 2862 |
+
"epoch": 9.15,
|
| 2863 |
+
"learning_rate": 4.230769230769231e-06,
|
| 2864 |
+
"loss": 2.1512,
|
| 2865 |
+
"step": 2856
|
| 2866 |
+
},
|
| 2867 |
+
{
|
| 2868 |
+
"epoch": 9.17,
|
| 2869 |
+
"learning_rate": 4.134615384615385e-06,
|
| 2870 |
+
"loss": 2.2936,
|
| 2871 |
+
"step": 2862
|
| 2872 |
+
},
|
| 2873 |
+
{
|
| 2874 |
+
"epoch": 9.19,
|
| 2875 |
+
"learning_rate": 4.0384615384615385e-06,
|
| 2876 |
+
"loss": 2.3936,
|
| 2877 |
+
"step": 2868
|
| 2878 |
+
},
|
| 2879 |
+
{
|
| 2880 |
+
"epoch": 9.21,
|
| 2881 |
+
"learning_rate": 3.942307692307692e-06,
|
| 2882 |
+
"loss": 2.1448,
|
| 2883 |
+
"step": 2874
|
| 2884 |
+
},
|
| 2885 |
+
{
|
| 2886 |
+
"epoch": 9.23,
|
| 2887 |
+
"learning_rate": 3.846153846153847e-06,
|
| 2888 |
+
"loss": 2.1278,
|
| 2889 |
+
"step": 2880
|
| 2890 |
+
},
|
| 2891 |
+
{
|
| 2892 |
+
"epoch": 9.25,
|
| 2893 |
+
"learning_rate": 3.75e-06,
|
| 2894 |
+
"loss": 2.2774,
|
| 2895 |
+
"step": 2886
|
| 2896 |
+
},
|
| 2897 |
+
{
|
| 2898 |
+
"epoch": 9.27,
|
| 2899 |
+
"learning_rate": 3.6538461538461542e-06,
|
| 2900 |
+
"loss": 2.1717,
|
| 2901 |
+
"step": 2892
|
| 2902 |
+
},
|
| 2903 |
+
{
|
| 2904 |
+
"epoch": 9.29,
|
| 2905 |
+
"learning_rate": 3.5576923076923075e-06,
|
| 2906 |
+
"loss": 2.3134,
|
| 2907 |
+
"step": 2898
|
| 2908 |
+
},
|
| 2909 |
+
{
|
| 2910 |
+
"epoch": 9.31,
|
| 2911 |
+
"learning_rate": 3.4615384615384617e-06,
|
| 2912 |
+
"loss": 2.0993,
|
| 2913 |
+
"step": 2904
|
| 2914 |
+
},
|
| 2915 |
+
{
|
| 2916 |
+
"epoch": 9.33,
|
| 2917 |
+
"learning_rate": 3.3653846153846154e-06,
|
| 2918 |
+
"loss": 2.1651,
|
| 2919 |
+
"step": 2910
|
| 2920 |
+
},
|
| 2921 |
+
{
|
| 2922 |
+
"epoch": 9.35,
|
| 2923 |
+
"learning_rate": 3.2692307692307696e-06,
|
| 2924 |
+
"loss": 2.2198,
|
| 2925 |
+
"step": 2916
|
| 2926 |
+
},
|
| 2927 |
+
{
|
| 2928 |
+
"epoch": 9.37,
|
| 2929 |
+
"learning_rate": 3.173076923076923e-06,
|
| 2930 |
+
"loss": 2.1949,
|
| 2931 |
+
"step": 2922
|
| 2932 |
+
},
|
| 2933 |
+
{
|
| 2934 |
+
"epoch": 9.38,
|
| 2935 |
+
"learning_rate": 3.0769230769230774e-06,
|
| 2936 |
+
"loss": 2.3332,
|
| 2937 |
+
"step": 2928
|
| 2938 |
+
},
|
| 2939 |
+
{
|
| 2940 |
+
"epoch": 9.4,
|
| 2941 |
+
"learning_rate": 2.980769230769231e-06,
|
| 2942 |
+
"loss": 2.175,
|
| 2943 |
+
"step": 2934
|
| 2944 |
+
},
|
| 2945 |
+
{
|
| 2946 |
+
"epoch": 9.42,
|
| 2947 |
+
"learning_rate": 2.884615384615385e-06,
|
| 2948 |
+
"loss": 2.2587,
|
| 2949 |
+
"step": 2940
|
| 2950 |
+
},
|
| 2951 |
+
{
|
| 2952 |
+
"epoch": 9.44,
|
| 2953 |
+
"learning_rate": 2.7884615384615386e-06,
|
| 2954 |
+
"loss": 2.1465,
|
| 2955 |
+
"step": 2946
|
| 2956 |
+
},
|
| 2957 |
+
{
|
| 2958 |
+
"epoch": 9.46,
|
| 2959 |
+
"learning_rate": 2.6923076923076928e-06,
|
| 2960 |
+
"loss": 2.1998,
|
| 2961 |
+
"step": 2952
|
| 2962 |
+
},
|
| 2963 |
+
{
|
| 2964 |
+
"epoch": 9.48,
|
| 2965 |
+
"learning_rate": 2.5961538461538465e-06,
|
| 2966 |
+
"loss": 2.2076,
|
| 2967 |
+
"step": 2958
|
| 2968 |
+
},
|
| 2969 |
+
{
|
| 2970 |
+
"epoch": 9.5,
|
| 2971 |
+
"learning_rate": 2.5e-06,
|
| 2972 |
+
"loss": 2.2025,
|
| 2973 |
+
"step": 2964
|
| 2974 |
+
},
|
| 2975 |
+
{
|
| 2976 |
+
"epoch": 9.52,
|
| 2977 |
+
"learning_rate": 2.403846153846154e-06,
|
| 2978 |
+
"loss": 2.1191,
|
| 2979 |
+
"step": 2970
|
| 2980 |
+
},
|
| 2981 |
+
{
|
| 2982 |
+
"epoch": 9.54,
|
| 2983 |
+
"learning_rate": 2.307692307692308e-06,
|
| 2984 |
+
"loss": 2.2118,
|
| 2985 |
+
"step": 2976
|
| 2986 |
+
},
|
| 2987 |
+
{
|
| 2988 |
+
"epoch": 9.56,
|
| 2989 |
+
"learning_rate": 2.211538461538462e-06,
|
| 2990 |
+
"loss": 2.1932,
|
| 2991 |
+
"step": 2982
|
| 2992 |
+
},
|
| 2993 |
+
{
|
| 2994 |
+
"epoch": 9.58,
|
| 2995 |
+
"learning_rate": 2.1153846153846155e-06,
|
| 2996 |
+
"loss": 2.2894,
|
| 2997 |
+
"step": 2988
|
| 2998 |
+
},
|
| 2999 |
+
{
|
| 3000 |
+
"epoch": 9.6,
|
| 3001 |
+
"learning_rate": 2.0192307692307692e-06,
|
| 3002 |
+
"loss": 2.3296,
|
| 3003 |
+
"step": 2994
|
| 3004 |
+
},
|
| 3005 |
+
{
|
| 3006 |
+
"epoch": 9.62,
|
| 3007 |
+
"learning_rate": 1.9230769230769234e-06,
|
| 3008 |
+
"loss": 2.2897,
|
| 3009 |
+
"step": 3000
|
| 3010 |
+
}
|
| 3011 |
+
],
|
| 3012 |
+
"logging_steps": 6,
|
| 3013 |
+
"max_steps": 3120,
|
| 3014 |
+
"num_train_epochs": 10,
|
| 3015 |
+
"save_steps": 500,
|
| 3016 |
+
"total_flos": 1208930867712000.0,
|
| 3017 |
+
"trial_name": null,
|
| 3018 |
+
"trial_params": null
|
| 3019 |
+
}
|
checkpoint-3000/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c150481b2be1a59fc2f8a97e3f83518bf0fe89d39fbfc7d3d90a4bdd3f30ccaf
|
| 3 |
+
size 4600
|
checkpoint-500/config.json
ADDED
|
@@ -0,0 +1,40 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "flax-community/gpt2-small-indonesian",
|
| 3 |
+
"activation_function": "gelu_new",
|
| 4 |
+
"architectures": [
|
| 5 |
+
"GPT2LMHeadModel"
|
| 6 |
+
],
|
| 7 |
+
"attn_pdrop": 0.0,
|
| 8 |
+
"bos_token_id": 50256,
|
| 9 |
+
"embd_pdrop": 0.0,
|
| 10 |
+
"eos_token_id": 50256,
|
| 11 |
+
"gradient_checkpointing": false,
|
| 12 |
+
"initializer_range": 0.02,
|
| 13 |
+
"layer_norm_epsilon": 1e-05,
|
| 14 |
+
"model_type": "gpt2",
|
| 15 |
+
"n_ctx": 1024,
|
| 16 |
+
"n_embd": 768,
|
| 17 |
+
"n_head": 12,
|
| 18 |
+
"n_inner": null,
|
| 19 |
+
"n_layer": 12,
|
| 20 |
+
"n_positions": 1024,
|
| 21 |
+
"reorder_and_upcast_attn": false,
|
| 22 |
+
"resid_pdrop": 0.0,
|
| 23 |
+
"scale_attn_by_inverse_layer_idx": false,
|
| 24 |
+
"scale_attn_weights": true,
|
| 25 |
+
"summary_activation": null,
|
| 26 |
+
"summary_first_dropout": 0.1,
|
| 27 |
+
"summary_proj_to_labels": true,
|
| 28 |
+
"summary_type": "cls_index",
|
| 29 |
+
"summary_use_proj": true,
|
| 30 |
+
"task_specific_params": {
|
| 31 |
+
"text-generation": {
|
| 32 |
+
"do_sample": true,
|
| 33 |
+
"max_length": 50
|
| 34 |
+
}
|
| 35 |
+
},
|
| 36 |
+
"torch_dtype": "float32",
|
| 37 |
+
"transformers_version": "4.35.2",
|
| 38 |
+
"use_cache": true,
|
| 39 |
+
"vocab_size": 50260
|
| 40 |
+
}
|
checkpoint-500/generation_config.json
ADDED
|
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"bos_token_id": 50256,
|
| 4 |
+
"eos_token_id": 50256,
|
| 5 |
+
"transformers_version": "4.35.2"
|
| 6 |
+
}
|
checkpoint-500/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:731c4a56c0fc8f5f8bc9b0c6a8f2103e73d50eb63c6fd203f274e0b7254537b9
|
| 3 |
+
size 497783424
|
checkpoint-500/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:fa6628a3d9fa7eb94776c99e61a4d6c3961f3d089b826505c4a9c3966dd547b3
|
| 3 |
+
size 995660730
|
checkpoint-500/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8d138cfe3a4adf21f048848ee35837c9a757a0a3616ff7adbb45b69aac247435
|
| 3 |
+
size 14244
|
checkpoint-500/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1742fcee019e207c02be6befeebdc0c74947e56ab8b3bf3f09e2c82d0388ebe5
|
| 3 |
+
size 1064
|
checkpoint-500/trainer_state.json
ADDED
|
@@ -0,0 +1,517 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 1.6025641025641026,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 500,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.02,
|
| 13 |
+
"learning_rate": 4.9903846153846154e-05,
|
| 14 |
+
"loss": 5.8218,
|
| 15 |
+
"step": 6
|
| 16 |
+
},
|
| 17 |
+
{
|
| 18 |
+
"epoch": 0.04,
|
| 19 |
+
"learning_rate": 4.980769230769231e-05,
|
| 20 |
+
"loss": 5.2499,
|
| 21 |
+
"step": 12
|
| 22 |
+
},
|
| 23 |
+
{
|
| 24 |
+
"epoch": 0.06,
|
| 25 |
+
"learning_rate": 4.9711538461538465e-05,
|
| 26 |
+
"loss": 5.1236,
|
| 27 |
+
"step": 18
|
| 28 |
+
},
|
| 29 |
+
{
|
| 30 |
+
"epoch": 0.08,
|
| 31 |
+
"learning_rate": 4.961538461538462e-05,
|
| 32 |
+
"loss": 4.8363,
|
| 33 |
+
"step": 24
|
| 34 |
+
},
|
| 35 |
+
{
|
| 36 |
+
"epoch": 0.1,
|
| 37 |
+
"learning_rate": 4.9519230769230776e-05,
|
| 38 |
+
"loss": 4.5805,
|
| 39 |
+
"step": 30
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.12,
|
| 43 |
+
"learning_rate": 4.942307692307693e-05,
|
| 44 |
+
"loss": 4.5276,
|
| 45 |
+
"step": 36
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.13,
|
| 49 |
+
"learning_rate": 4.932692307692308e-05,
|
| 50 |
+
"loss": 4.3871,
|
| 51 |
+
"step": 42
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.15,
|
| 55 |
+
"learning_rate": 4.923076923076924e-05,
|
| 56 |
+
"loss": 4.2706,
|
| 57 |
+
"step": 48
|
| 58 |
+
},
|
| 59 |
+
{
|
| 60 |
+
"epoch": 0.17,
|
| 61 |
+
"learning_rate": 4.9134615384615384e-05,
|
| 62 |
+
"loss": 4.0906,
|
| 63 |
+
"step": 54
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 0.19,
|
| 67 |
+
"learning_rate": 4.9038461538461536e-05,
|
| 68 |
+
"loss": 4.1704,
|
| 69 |
+
"step": 60
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 0.21,
|
| 73 |
+
"learning_rate": 4.8942307692307695e-05,
|
| 74 |
+
"loss": 3.9014,
|
| 75 |
+
"step": 66
|
| 76 |
+
},
|
| 77 |
+
{
|
| 78 |
+
"epoch": 0.23,
|
| 79 |
+
"learning_rate": 4.884615384615385e-05,
|
| 80 |
+
"loss": 4.0338,
|
| 81 |
+
"step": 72
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"epoch": 0.25,
|
| 85 |
+
"learning_rate": 4.875e-05,
|
| 86 |
+
"loss": 4.011,
|
| 87 |
+
"step": 78
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"epoch": 0.27,
|
| 91 |
+
"learning_rate": 4.865384615384616e-05,
|
| 92 |
+
"loss": 3.8017,
|
| 93 |
+
"step": 84
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.29,
|
| 97 |
+
"learning_rate": 4.855769230769231e-05,
|
| 98 |
+
"loss": 3.819,
|
| 99 |
+
"step": 90
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"epoch": 0.31,
|
| 103 |
+
"learning_rate": 4.846153846153846e-05,
|
| 104 |
+
"loss": 3.8157,
|
| 105 |
+
"step": 96
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 0.33,
|
| 109 |
+
"learning_rate": 4.836538461538462e-05,
|
| 110 |
+
"loss": 3.7675,
|
| 111 |
+
"step": 102
|
| 112 |
+
},
|
| 113 |
+
{
|
| 114 |
+
"epoch": 0.35,
|
| 115 |
+
"learning_rate": 4.826923076923077e-05,
|
| 116 |
+
"loss": 3.756,
|
| 117 |
+
"step": 108
|
| 118 |
+
},
|
| 119 |
+
{
|
| 120 |
+
"epoch": 0.37,
|
| 121 |
+
"learning_rate": 4.8173076923076925e-05,
|
| 122 |
+
"loss": 3.7439,
|
| 123 |
+
"step": 114
|
| 124 |
+
},
|
| 125 |
+
{
|
| 126 |
+
"epoch": 0.38,
|
| 127 |
+
"learning_rate": 4.8076923076923084e-05,
|
| 128 |
+
"loss": 3.761,
|
| 129 |
+
"step": 120
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.4,
|
| 133 |
+
"learning_rate": 4.7980769230769236e-05,
|
| 134 |
+
"loss": 3.7318,
|
| 135 |
+
"step": 126
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.42,
|
| 139 |
+
"learning_rate": 4.788461538461539e-05,
|
| 140 |
+
"loss": 3.745,
|
| 141 |
+
"step": 132
|
| 142 |
+
},
|
| 143 |
+
{
|
| 144 |
+
"epoch": 0.44,
|
| 145 |
+
"learning_rate": 4.778846153846154e-05,
|
| 146 |
+
"loss": 3.784,
|
| 147 |
+
"step": 138
|
| 148 |
+
},
|
| 149 |
+
{
|
| 150 |
+
"epoch": 0.46,
|
| 151 |
+
"learning_rate": 4.76923076923077e-05,
|
| 152 |
+
"loss": 3.6114,
|
| 153 |
+
"step": 144
|
| 154 |
+
},
|
| 155 |
+
{
|
| 156 |
+
"epoch": 0.48,
|
| 157 |
+
"learning_rate": 4.7596153846153844e-05,
|
| 158 |
+
"loss": 3.637,
|
| 159 |
+
"step": 150
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 0.5,
|
| 163 |
+
"learning_rate": 4.75e-05,
|
| 164 |
+
"loss": 3.5881,
|
| 165 |
+
"step": 156
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 0.52,
|
| 169 |
+
"learning_rate": 4.7403846153846155e-05,
|
| 170 |
+
"loss": 3.5336,
|
| 171 |
+
"step": 162
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 0.54,
|
| 175 |
+
"learning_rate": 4.730769230769231e-05,
|
| 176 |
+
"loss": 3.5313,
|
| 177 |
+
"step": 168
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.56,
|
| 181 |
+
"learning_rate": 4.7211538461538465e-05,
|
| 182 |
+
"loss": 3.6357,
|
| 183 |
+
"step": 174
|
| 184 |
+
},
|
| 185 |
+
{
|
| 186 |
+
"epoch": 0.58,
|
| 187 |
+
"learning_rate": 4.711538461538462e-05,
|
| 188 |
+
"loss": 3.5818,
|
| 189 |
+
"step": 180
|
| 190 |
+
},
|
| 191 |
+
{
|
| 192 |
+
"epoch": 0.6,
|
| 193 |
+
"learning_rate": 4.701923076923077e-05,
|
| 194 |
+
"loss": 3.5017,
|
| 195 |
+
"step": 186
|
| 196 |
+
},
|
| 197 |
+
{
|
| 198 |
+
"epoch": 0.62,
|
| 199 |
+
"learning_rate": 4.692307692307693e-05,
|
| 200 |
+
"loss": 3.5815,
|
| 201 |
+
"step": 192
|
| 202 |
+
},
|
| 203 |
+
{
|
| 204 |
+
"epoch": 0.63,
|
| 205 |
+
"learning_rate": 4.682692307692308e-05,
|
| 206 |
+
"loss": 3.611,
|
| 207 |
+
"step": 198
|
| 208 |
+
},
|
| 209 |
+
{
|
| 210 |
+
"epoch": 0.65,
|
| 211 |
+
"learning_rate": 4.673076923076923e-05,
|
| 212 |
+
"loss": 3.6096,
|
| 213 |
+
"step": 204
|
| 214 |
+
},
|
| 215 |
+
{
|
| 216 |
+
"epoch": 0.67,
|
| 217 |
+
"learning_rate": 4.6634615384615384e-05,
|
| 218 |
+
"loss": 3.5436,
|
| 219 |
+
"step": 210
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.69,
|
| 223 |
+
"learning_rate": 4.653846153846154e-05,
|
| 224 |
+
"loss": 3.5258,
|
| 225 |
+
"step": 216
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 0.71,
|
| 229 |
+
"learning_rate": 4.6442307692307695e-05,
|
| 230 |
+
"loss": 3.5277,
|
| 231 |
+
"step": 222
|
| 232 |
+
},
|
| 233 |
+
{
|
| 234 |
+
"epoch": 0.73,
|
| 235 |
+
"learning_rate": 4.634615384615385e-05,
|
| 236 |
+
"loss": 3.505,
|
| 237 |
+
"step": 228
|
| 238 |
+
},
|
| 239 |
+
{
|
| 240 |
+
"epoch": 0.75,
|
| 241 |
+
"learning_rate": 4.6250000000000006e-05,
|
| 242 |
+
"loss": 3.4665,
|
| 243 |
+
"step": 234
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"epoch": 0.77,
|
| 247 |
+
"learning_rate": 4.615384615384616e-05,
|
| 248 |
+
"loss": 3.4245,
|
| 249 |
+
"step": 240
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 0.79,
|
| 253 |
+
"learning_rate": 4.605769230769231e-05,
|
| 254 |
+
"loss": 3.4547,
|
| 255 |
+
"step": 246
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 0.81,
|
| 259 |
+
"learning_rate": 4.596153846153846e-05,
|
| 260 |
+
"loss": 3.4382,
|
| 261 |
+
"step": 252
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.83,
|
| 265 |
+
"learning_rate": 4.5865384615384614e-05,
|
| 266 |
+
"loss": 3.482,
|
| 267 |
+
"step": 258
|
| 268 |
+
},
|
| 269 |
+
{
|
| 270 |
+
"epoch": 0.85,
|
| 271 |
+
"learning_rate": 4.576923076923077e-05,
|
| 272 |
+
"loss": 3.4795,
|
| 273 |
+
"step": 264
|
| 274 |
+
},
|
| 275 |
+
{
|
| 276 |
+
"epoch": 0.87,
|
| 277 |
+
"learning_rate": 4.5673076923076925e-05,
|
| 278 |
+
"loss": 3.5018,
|
| 279 |
+
"step": 270
|
| 280 |
+
},
|
| 281 |
+
{
|
| 282 |
+
"epoch": 0.88,
|
| 283 |
+
"learning_rate": 4.557692307692308e-05,
|
| 284 |
+
"loss": 3.448,
|
| 285 |
+
"step": 276
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 0.9,
|
| 289 |
+
"learning_rate": 4.548076923076923e-05,
|
| 290 |
+
"loss": 3.4206,
|
| 291 |
+
"step": 282
|
| 292 |
+
},
|
| 293 |
+
{
|
| 294 |
+
"epoch": 0.92,
|
| 295 |
+
"learning_rate": 4.538461538461539e-05,
|
| 296 |
+
"loss": 3.528,
|
| 297 |
+
"step": 288
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 0.94,
|
| 301 |
+
"learning_rate": 4.528846153846154e-05,
|
| 302 |
+
"loss": 3.475,
|
| 303 |
+
"step": 294
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.96,
|
| 307 |
+
"learning_rate": 4.519230769230769e-05,
|
| 308 |
+
"loss": 3.4142,
|
| 309 |
+
"step": 300
|
| 310 |
+
},
|
| 311 |
+
{
|
| 312 |
+
"epoch": 0.98,
|
| 313 |
+
"learning_rate": 4.509615384615385e-05,
|
| 314 |
+
"loss": 3.4156,
|
| 315 |
+
"step": 306
|
| 316 |
+
},
|
| 317 |
+
{
|
| 318 |
+
"epoch": 1.0,
|
| 319 |
+
"learning_rate": 4.5e-05,
|
| 320 |
+
"loss": 3.3274,
|
| 321 |
+
"step": 312
|
| 322 |
+
},
|
| 323 |
+
{
|
| 324 |
+
"epoch": 1.02,
|
| 325 |
+
"learning_rate": 4.4903846153846155e-05,
|
| 326 |
+
"loss": 3.3594,
|
| 327 |
+
"step": 318
|
| 328 |
+
},
|
| 329 |
+
{
|
| 330 |
+
"epoch": 1.04,
|
| 331 |
+
"learning_rate": 4.4807692307692314e-05,
|
| 332 |
+
"loss": 3.2517,
|
| 333 |
+
"step": 324
|
| 334 |
+
},
|
| 335 |
+
{
|
| 336 |
+
"epoch": 1.06,
|
| 337 |
+
"learning_rate": 4.4711538461538466e-05,
|
| 338 |
+
"loss": 3.2934,
|
| 339 |
+
"step": 330
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 1.08,
|
| 343 |
+
"learning_rate": 4.461538461538462e-05,
|
| 344 |
+
"loss": 3.2986,
|
| 345 |
+
"step": 336
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 1.1,
|
| 349 |
+
"learning_rate": 4.451923076923077e-05,
|
| 350 |
+
"loss": 3.2655,
|
| 351 |
+
"step": 342
|
| 352 |
+
},
|
| 353 |
+
{
|
| 354 |
+
"epoch": 1.12,
|
| 355 |
+
"learning_rate": 4.442307692307692e-05,
|
| 356 |
+
"loss": 3.4058,
|
| 357 |
+
"step": 348
|
| 358 |
+
},
|
| 359 |
+
{
|
| 360 |
+
"epoch": 1.13,
|
| 361 |
+
"learning_rate": 4.4326923076923074e-05,
|
| 362 |
+
"loss": 3.265,
|
| 363 |
+
"step": 354
|
| 364 |
+
},
|
| 365 |
+
{
|
| 366 |
+
"epoch": 1.15,
|
| 367 |
+
"learning_rate": 4.423076923076923e-05,
|
| 368 |
+
"loss": 3.3208,
|
| 369 |
+
"step": 360
|
| 370 |
+
},
|
| 371 |
+
{
|
| 372 |
+
"epoch": 1.17,
|
| 373 |
+
"learning_rate": 4.4134615384615385e-05,
|
| 374 |
+
"loss": 3.238,
|
| 375 |
+
"step": 366
|
| 376 |
+
},
|
| 377 |
+
{
|
| 378 |
+
"epoch": 1.19,
|
| 379 |
+
"learning_rate": 4.403846153846154e-05,
|
| 380 |
+
"loss": 3.2153,
|
| 381 |
+
"step": 372
|
| 382 |
+
},
|
| 383 |
+
{
|
| 384 |
+
"epoch": 1.21,
|
| 385 |
+
"learning_rate": 4.3942307692307695e-05,
|
| 386 |
+
"loss": 3.264,
|
| 387 |
+
"step": 378
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 1.23,
|
| 391 |
+
"learning_rate": 4.384615384615385e-05,
|
| 392 |
+
"loss": 3.2911,
|
| 393 |
+
"step": 384
|
| 394 |
+
},
|
| 395 |
+
{
|
| 396 |
+
"epoch": 1.25,
|
| 397 |
+
"learning_rate": 4.375e-05,
|
| 398 |
+
"loss": 3.3027,
|
| 399 |
+
"step": 390
|
| 400 |
+
},
|
| 401 |
+
{
|
| 402 |
+
"epoch": 1.27,
|
| 403 |
+
"learning_rate": 4.365384615384616e-05,
|
| 404 |
+
"loss": 3.2589,
|
| 405 |
+
"step": 396
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 1.29,
|
| 409 |
+
"learning_rate": 4.355769230769231e-05,
|
| 410 |
+
"loss": 3.3683,
|
| 411 |
+
"step": 402
|
| 412 |
+
},
|
| 413 |
+
{
|
| 414 |
+
"epoch": 1.31,
|
| 415 |
+
"learning_rate": 4.346153846153846e-05,
|
| 416 |
+
"loss": 3.2849,
|
| 417 |
+
"step": 408
|
| 418 |
+
},
|
| 419 |
+
{
|
| 420 |
+
"epoch": 1.33,
|
| 421 |
+
"learning_rate": 4.336538461538462e-05,
|
| 422 |
+
"loss": 3.2397,
|
| 423 |
+
"step": 414
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 1.35,
|
| 427 |
+
"learning_rate": 4.326923076923077e-05,
|
| 428 |
+
"loss": 3.2128,
|
| 429 |
+
"step": 420
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 1.37,
|
| 433 |
+
"learning_rate": 4.3173076923076925e-05,
|
| 434 |
+
"loss": 3.1944,
|
| 435 |
+
"step": 426
|
| 436 |
+
},
|
| 437 |
+
{
|
| 438 |
+
"epoch": 1.38,
|
| 439 |
+
"learning_rate": 4.3076923076923084e-05,
|
| 440 |
+
"loss": 3.1837,
|
| 441 |
+
"step": 432
|
| 442 |
+
},
|
| 443 |
+
{
|
| 444 |
+
"epoch": 1.4,
|
| 445 |
+
"learning_rate": 4.2980769230769236e-05,
|
| 446 |
+
"loss": 3.1793,
|
| 447 |
+
"step": 438
|
| 448 |
+
},
|
| 449 |
+
{
|
| 450 |
+
"epoch": 1.42,
|
| 451 |
+
"learning_rate": 4.288461538461538e-05,
|
| 452 |
+
"loss": 3.1447,
|
| 453 |
+
"step": 444
|
| 454 |
+
},
|
| 455 |
+
{
|
| 456 |
+
"epoch": 1.44,
|
| 457 |
+
"learning_rate": 4.278846153846154e-05,
|
| 458 |
+
"loss": 3.1028,
|
| 459 |
+
"step": 450
|
| 460 |
+
},
|
| 461 |
+
{
|
| 462 |
+
"epoch": 1.46,
|
| 463 |
+
"learning_rate": 4.269230769230769e-05,
|
| 464 |
+
"loss": 3.2471,
|
| 465 |
+
"step": 456
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 1.48,
|
| 469 |
+
"learning_rate": 4.2596153846153844e-05,
|
| 470 |
+
"loss": 3.1855,
|
| 471 |
+
"step": 462
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 1.5,
|
| 475 |
+
"learning_rate": 4.25e-05,
|
| 476 |
+
"loss": 3.1817,
|
| 477 |
+
"step": 468
|
| 478 |
+
},
|
| 479 |
+
{
|
| 480 |
+
"epoch": 1.52,
|
| 481 |
+
"learning_rate": 4.2403846153846155e-05,
|
| 482 |
+
"loss": 3.214,
|
| 483 |
+
"step": 474
|
| 484 |
+
},
|
| 485 |
+
{
|
| 486 |
+
"epoch": 1.54,
|
| 487 |
+
"learning_rate": 4.230769230769231e-05,
|
| 488 |
+
"loss": 2.997,
|
| 489 |
+
"step": 480
|
| 490 |
+
},
|
| 491 |
+
{
|
| 492 |
+
"epoch": 1.56,
|
| 493 |
+
"learning_rate": 4.2211538461538466e-05,
|
| 494 |
+
"loss": 3.2059,
|
| 495 |
+
"step": 486
|
| 496 |
+
},
|
| 497 |
+
{
|
| 498 |
+
"epoch": 1.58,
|
| 499 |
+
"learning_rate": 4.211538461538462e-05,
|
| 500 |
+
"loss": 3.1517,
|
| 501 |
+
"step": 492
|
| 502 |
+
},
|
| 503 |
+
{
|
| 504 |
+
"epoch": 1.6,
|
| 505 |
+
"learning_rate": 4.201923076923077e-05,
|
| 506 |
+
"loss": 3.2236,
|
| 507 |
+
"step": 498
|
| 508 |
+
}
|
| 509 |
+
],
|
| 510 |
+
"logging_steps": 6,
|
| 511 |
+
"max_steps": 3120,
|
| 512 |
+
"num_train_epochs": 10,
|
| 513 |
+
"save_steps": 500,
|
| 514 |
+
"total_flos": 201131582976000.0,
|
| 515 |
+
"trial_name": null,
|
| 516 |
+
"trial_params": null
|
| 517 |
+
}
|
checkpoint-500/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c150481b2be1a59fc2f8a97e3f83518bf0fe89d39fbfc7d3d90a4bdd3f30ccaf
|
| 3 |
+
size 4600
|
config.json
ADDED
|
@@ -0,0 +1,40 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "flax-community/gpt2-small-indonesian",
|
| 3 |
+
"activation_function": "gelu_new",
|
| 4 |
+
"architectures": [
|
| 5 |
+
"GPT2LMHeadModel"
|
| 6 |
+
],
|
| 7 |
+
"attn_pdrop": 0.0,
|
| 8 |
+
"bos_token_id": 50256,
|
| 9 |
+
"embd_pdrop": 0.0,
|
| 10 |
+
"eos_token_id": 50256,
|
| 11 |
+
"gradient_checkpointing": false,
|
| 12 |
+
"initializer_range": 0.02,
|
| 13 |
+
"layer_norm_epsilon": 1e-05,
|
| 14 |
+
"model_type": "gpt2",
|
| 15 |
+
"n_ctx": 1024,
|
| 16 |
+
"n_embd": 768,
|
| 17 |
+
"n_head": 12,
|
| 18 |
+
"n_inner": null,
|
| 19 |
+
"n_layer": 12,
|
| 20 |
+
"n_positions": 1024,
|
| 21 |
+
"reorder_and_upcast_attn": false,
|
| 22 |
+
"resid_pdrop": 0.0,
|
| 23 |
+
"scale_attn_by_inverse_layer_idx": false,
|
| 24 |
+
"scale_attn_weights": true,
|
| 25 |
+
"summary_activation": null,
|
| 26 |
+
"summary_first_dropout": 0.1,
|
| 27 |
+
"summary_proj_to_labels": true,
|
| 28 |
+
"summary_type": "cls_index",
|
| 29 |
+
"summary_use_proj": true,
|
| 30 |
+
"task_specific_params": {
|
| 31 |
+
"text-generation": {
|
| 32 |
+
"do_sample": true,
|
| 33 |
+
"max_length": 50
|
| 34 |
+
}
|
| 35 |
+
},
|
| 36 |
+
"torch_dtype": "float32",
|
| 37 |
+
"transformers_version": "4.35.2",
|
| 38 |
+
"use_cache": true,
|
| 39 |
+
"vocab_size": 50260
|
| 40 |
+
}
|