Upload 48 files
Browse files- checkpoint-10000/added_tokens.json +3 -0
- checkpoint-10000/optimizer.pt +3 -0
- checkpoint-10000/pytorch_model.bin +3 -0
- checkpoint-10000/rng_state.pth +3 -0
- checkpoint-10000/scaler.pt +3 -0
- checkpoint-10000/scheduler.pt +3 -0
- checkpoint-10000/special_tokens_map.json +24 -0
- checkpoint-10000/tokenizer.json +0 -0
- checkpoint-10000/tokenizer.model +3 -0
- checkpoint-10000/tokenizer_config.json +34 -0
- checkpoint-10000/trainer_state.json +624 -0
- checkpoint-10000/training_args.bin +3 -0
- checkpoint-20000/added_tokens.json +3 -0
- checkpoint-20000/optimizer.pt +3 -0
- checkpoint-20000/pytorch_model.bin +3 -0
- checkpoint-20000/rng_state.pth +3 -0
- checkpoint-20000/scaler.pt +3 -0
- checkpoint-20000/scheduler.pt +3 -0
- checkpoint-20000/special_tokens_map.json +24 -0
- checkpoint-20000/tokenizer.json +0 -0
- checkpoint-20000/tokenizer.model +3 -0
- checkpoint-20000/tokenizer_config.json +34 -0
- checkpoint-20000/trainer_state.json +1232 -0
- checkpoint-20000/training_args.bin +3 -0
- checkpoint-30000/added_tokens.json +3 -0
- checkpoint-30000/optimizer.pt +3 -0
- checkpoint-30000/pytorch_model.bin +3 -0
- checkpoint-30000/rng_state.pth +3 -0
- checkpoint-30000/scaler.pt +3 -0
- checkpoint-30000/scheduler.pt +3 -0
- checkpoint-30000/special_tokens_map.json +24 -0
- checkpoint-30000/tokenizer.json +0 -0
- checkpoint-30000/tokenizer.model +3 -0
- checkpoint-30000/tokenizer_config.json +34 -0
- checkpoint-30000/trainer_state.json +1840 -0
- checkpoint-30000/training_args.bin +3 -0
- checkpoint-40000/added_tokens.json +3 -0
- checkpoint-40000/optimizer.pt +3 -0
- checkpoint-40000/pytorch_model.bin +3 -0
- checkpoint-40000/rng_state.pth +3 -0
- checkpoint-40000/scaler.pt +3 -0
- checkpoint-40000/scheduler.pt +3 -0
- checkpoint-40000/special_tokens_map.json +24 -0
- checkpoint-40000/tokenizer.json +0 -0
- checkpoint-40000/tokenizer.model +3 -0
- checkpoint-40000/tokenizer_config.json +34 -0
- checkpoint-40000/trainer_state.json +2448 -0
- checkpoint-40000/training_args.bin +3 -0
checkpoint-10000/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<pad>": 32000
|
| 3 |
+
}
|
checkpoint-10000/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4117c112df6b7956bdd7b7196a762ae684a093e84309829fe479118fb3812386
|
| 3 |
+
size 134433530
|
checkpoint-10000/pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:db09a2c51f9a3ef13f3b100ed6cb9f2ddeeb71b05c166e5ef6ce3b075d1c8853
|
| 3 |
+
size 67201802
|
checkpoint-10000/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:179e8e5a0c5e195b1d4035c58ac618e70b8cd3cfd6b54916d678c105bfd9b34b
|
| 3 |
+
size 14244
|
checkpoint-10000/scaler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:64171e50e71471042fc95724b4b2d26c212d3ee1e2d7063b4ecee7db45996ca6
|
| 3 |
+
size 988
|
checkpoint-10000/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:89a64efe49535d1cf6703f1d5ed7abe2d58cceadf9db33d1942a33af6f5d6084
|
| 3 |
+
size 1064
|
checkpoint-10000/special_tokens_map.json
ADDED
|
@@ -0,0 +1,24 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"content": "<s>",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": false,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"eos_token": {
|
| 10 |
+
"content": "</s>",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"pad_token": "<s>",
|
| 17 |
+
"unk_token": {
|
| 18 |
+
"content": "<unk>",
|
| 19 |
+
"lstrip": false,
|
| 20 |
+
"normalized": false,
|
| 21 |
+
"rstrip": false,
|
| 22 |
+
"single_word": false
|
| 23 |
+
}
|
| 24 |
+
}
|
checkpoint-10000/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
checkpoint-10000/tokenizer.model
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9e556afd44213b6bd1be2b850ebbbd98f5481437a8021afaf58ee7fb1818d347
|
| 3 |
+
size 499723
|
checkpoint-10000/tokenizer_config.json
ADDED
|
@@ -0,0 +1,34 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_bos_token": true,
|
| 3 |
+
"add_eos_token": false,
|
| 4 |
+
"bos_token": {
|
| 5 |
+
"__type": "AddedToken",
|
| 6 |
+
"content": "<s>",
|
| 7 |
+
"lstrip": false,
|
| 8 |
+
"normalized": false,
|
| 9 |
+
"rstrip": false,
|
| 10 |
+
"single_word": false
|
| 11 |
+
},
|
| 12 |
+
"clean_up_tokenization_spaces": false,
|
| 13 |
+
"eos_token": {
|
| 14 |
+
"__type": "AddedToken",
|
| 15 |
+
"content": "</s>",
|
| 16 |
+
"lstrip": false,
|
| 17 |
+
"normalized": false,
|
| 18 |
+
"rstrip": false,
|
| 19 |
+
"single_word": false
|
| 20 |
+
},
|
| 21 |
+
"legacy": false,
|
| 22 |
+
"model_max_length": 1000000000000000019884624838656,
|
| 23 |
+
"pad_token": null,
|
| 24 |
+
"sp_model_kwargs": {},
|
| 25 |
+
"tokenizer_class": "LlamaTokenizer",
|
| 26 |
+
"unk_token": {
|
| 27 |
+
"__type": "AddedToken",
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false
|
| 33 |
+
}
|
| 34 |
+
}
|
checkpoint-10000/trainer_state.json
ADDED
|
@@ -0,0 +1,624 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.740916907787323,
|
| 3 |
+
"best_model_checkpoint": "/media/xzuyn/c41d93ed-83a7-4577-9d38-5062b7e14438/xzuyn/Documents/AI_ROCM_5-6/LLaMa-LoRA-Tuner/data/lora_models/LLaMa-2-WizardLM-Uncensored-Tulu-Format-50K-7B-LoRA/checkpoint-10000",
|
| 4 |
+
"epoch": 0.2,
|
| 5 |
+
"global_step": 10000,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.0,
|
| 12 |
+
"learning_rate": 0.0001,
|
| 13 |
+
"loss": 1.1136,
|
| 14 |
+
"step": 100
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.0,
|
| 18 |
+
"learning_rate": 9.997999599919985e-05,
|
| 19 |
+
"loss": 0.9595,
|
| 20 |
+
"step": 200
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.01,
|
| 24 |
+
"learning_rate": 9.995999199839968e-05,
|
| 25 |
+
"loss": 0.8936,
|
| 26 |
+
"step": 300
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.01,
|
| 30 |
+
"learning_rate": 9.993998799759952e-05,
|
| 31 |
+
"loss": 0.9237,
|
| 32 |
+
"step": 400
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.01,
|
| 36 |
+
"learning_rate": 9.992038407681536e-05,
|
| 37 |
+
"loss": 0.9005,
|
| 38 |
+
"step": 500
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.01,
|
| 42 |
+
"learning_rate": 9.99003800760152e-05,
|
| 43 |
+
"loss": 0.8433,
|
| 44 |
+
"step": 600
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.01,
|
| 48 |
+
"learning_rate": 9.988037607521505e-05,
|
| 49 |
+
"loss": 0.863,
|
| 50 |
+
"step": 700
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.02,
|
| 54 |
+
"learning_rate": 9.98603720744149e-05,
|
| 55 |
+
"loss": 0.8634,
|
| 56 |
+
"step": 800
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.02,
|
| 60 |
+
"learning_rate": 9.984036807361474e-05,
|
| 61 |
+
"loss": 0.8463,
|
| 62 |
+
"step": 900
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.02,
|
| 66 |
+
"learning_rate": 9.982036407281457e-05,
|
| 67 |
+
"loss": 0.8527,
|
| 68 |
+
"step": 1000
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.02,
|
| 72 |
+
"learning_rate": 9.980036007201441e-05,
|
| 73 |
+
"loss": 0.8259,
|
| 74 |
+
"step": 1100
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.02,
|
| 78 |
+
"learning_rate": 9.978035607121424e-05,
|
| 79 |
+
"loss": 0.8473,
|
| 80 |
+
"step": 1200
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.03,
|
| 84 |
+
"learning_rate": 9.976035207041409e-05,
|
| 85 |
+
"loss": 0.8462,
|
| 86 |
+
"step": 1300
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.03,
|
| 90 |
+
"learning_rate": 9.974034806961393e-05,
|
| 91 |
+
"loss": 0.8609,
|
| 92 |
+
"step": 1400
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.03,
|
| 96 |
+
"learning_rate": 9.972034406881376e-05,
|
| 97 |
+
"loss": 0.8238,
|
| 98 |
+
"step": 1500
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.03,
|
| 102 |
+
"learning_rate": 9.970034006801361e-05,
|
| 103 |
+
"loss": 0.8212,
|
| 104 |
+
"step": 1600
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 0.03,
|
| 108 |
+
"learning_rate": 9.968033606721344e-05,
|
| 109 |
+
"loss": 0.8391,
|
| 110 |
+
"step": 1700
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.04,
|
| 114 |
+
"learning_rate": 9.966033206641328e-05,
|
| 115 |
+
"loss": 0.774,
|
| 116 |
+
"step": 1800
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 0.04,
|
| 120 |
+
"learning_rate": 9.964032806561313e-05,
|
| 121 |
+
"loss": 0.8149,
|
| 122 |
+
"step": 1900
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.04,
|
| 126 |
+
"learning_rate": 9.962032406481296e-05,
|
| 127 |
+
"loss": 0.8794,
|
| 128 |
+
"step": 2000
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.04,
|
| 132 |
+
"learning_rate": 9.96003200640128e-05,
|
| 133 |
+
"loss": 0.7747,
|
| 134 |
+
"step": 2100
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.04,
|
| 138 |
+
"learning_rate": 9.958031606321264e-05,
|
| 139 |
+
"loss": 0.8191,
|
| 140 |
+
"step": 2200
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 0.05,
|
| 144 |
+
"learning_rate": 9.956031206241248e-05,
|
| 145 |
+
"loss": 0.8219,
|
| 146 |
+
"step": 2300
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 0.05,
|
| 150 |
+
"learning_rate": 9.954030806161232e-05,
|
| 151 |
+
"loss": 0.8191,
|
| 152 |
+
"step": 2400
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 0.05,
|
| 156 |
+
"learning_rate": 9.952030406081217e-05,
|
| 157 |
+
"loss": 0.8662,
|
| 158 |
+
"step": 2500
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 0.05,
|
| 162 |
+
"learning_rate": 9.950030006001201e-05,
|
| 163 |
+
"loss": 0.787,
|
| 164 |
+
"step": 2600
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.05,
|
| 168 |
+
"learning_rate": 9.948029605921184e-05,
|
| 169 |
+
"loss": 0.8668,
|
| 170 |
+
"step": 2700
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.06,
|
| 174 |
+
"learning_rate": 9.946029205841169e-05,
|
| 175 |
+
"loss": 0.7817,
|
| 176 |
+
"step": 2800
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 0.06,
|
| 180 |
+
"learning_rate": 9.944028805761153e-05,
|
| 181 |
+
"loss": 0.8021,
|
| 182 |
+
"step": 2900
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 0.06,
|
| 186 |
+
"learning_rate": 9.942028405681136e-05,
|
| 187 |
+
"loss": 0.8324,
|
| 188 |
+
"step": 3000
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 0.06,
|
| 192 |
+
"learning_rate": 9.940028005601121e-05,
|
| 193 |
+
"loss": 0.8631,
|
| 194 |
+
"step": 3100
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 0.06,
|
| 198 |
+
"learning_rate": 9.938027605521105e-05,
|
| 199 |
+
"loss": 0.797,
|
| 200 |
+
"step": 3200
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 0.07,
|
| 204 |
+
"learning_rate": 9.936027205441088e-05,
|
| 205 |
+
"loss": 0.8224,
|
| 206 |
+
"step": 3300
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.07,
|
| 210 |
+
"learning_rate": 9.934026805361073e-05,
|
| 211 |
+
"loss": 0.8164,
|
| 212 |
+
"step": 3400
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.07,
|
| 216 |
+
"learning_rate": 9.932026405281056e-05,
|
| 217 |
+
"loss": 0.7975,
|
| 218 |
+
"step": 3500
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 0.07,
|
| 222 |
+
"learning_rate": 9.930046009201841e-05,
|
| 223 |
+
"loss": 0.8468,
|
| 224 |
+
"step": 3600
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 0.07,
|
| 228 |
+
"learning_rate": 9.928045609121824e-05,
|
| 229 |
+
"loss": 0.8169,
|
| 230 |
+
"step": 3700
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 0.08,
|
| 234 |
+
"learning_rate": 9.926045209041809e-05,
|
| 235 |
+
"loss": 0.7761,
|
| 236 |
+
"step": 3800
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 0.08,
|
| 240 |
+
"learning_rate": 9.924044808961793e-05,
|
| 241 |
+
"loss": 0.8042,
|
| 242 |
+
"step": 3900
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 0.08,
|
| 246 |
+
"learning_rate": 9.922044408881776e-05,
|
| 247 |
+
"loss": 0.8146,
|
| 248 |
+
"step": 4000
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 0.08,
|
| 252 |
+
"learning_rate": 9.920044008801761e-05,
|
| 253 |
+
"loss": 0.8243,
|
| 254 |
+
"step": 4100
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.08,
|
| 258 |
+
"learning_rate": 9.918043608721744e-05,
|
| 259 |
+
"loss": 0.8315,
|
| 260 |
+
"step": 4200
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 0.09,
|
| 264 |
+
"learning_rate": 9.916043208641728e-05,
|
| 265 |
+
"loss": 0.8014,
|
| 266 |
+
"step": 4300
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 0.09,
|
| 270 |
+
"learning_rate": 9.914042808561713e-05,
|
| 271 |
+
"loss": 0.8241,
|
| 272 |
+
"step": 4400
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 0.09,
|
| 276 |
+
"learning_rate": 9.912042408481696e-05,
|
| 277 |
+
"loss": 0.8233,
|
| 278 |
+
"step": 4500
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 0.09,
|
| 282 |
+
"learning_rate": 9.91004200840168e-05,
|
| 283 |
+
"loss": 0.7748,
|
| 284 |
+
"step": 4600
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 0.09,
|
| 288 |
+
"learning_rate": 9.908041608321665e-05,
|
| 289 |
+
"loss": 0.8197,
|
| 290 |
+
"step": 4700
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 0.1,
|
| 294 |
+
"learning_rate": 9.906041208241649e-05,
|
| 295 |
+
"loss": 0.8195,
|
| 296 |
+
"step": 4800
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.1,
|
| 300 |
+
"learning_rate": 9.904040808161634e-05,
|
| 301 |
+
"loss": 0.7864,
|
| 302 |
+
"step": 4900
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 0.1,
|
| 306 |
+
"learning_rate": 9.902040408081617e-05,
|
| 307 |
+
"loss": 0.7812,
|
| 308 |
+
"step": 5000
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 0.1,
|
| 312 |
+
"learning_rate": 9.900040008001601e-05,
|
| 313 |
+
"loss": 0.7966,
|
| 314 |
+
"step": 5100
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 0.1,
|
| 318 |
+
"learning_rate": 9.898039607921584e-05,
|
| 319 |
+
"loss": 0.793,
|
| 320 |
+
"step": 5200
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 0.11,
|
| 324 |
+
"learning_rate": 9.896039207841569e-05,
|
| 325 |
+
"loss": 0.8568,
|
| 326 |
+
"step": 5300
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 0.11,
|
| 330 |
+
"learning_rate": 9.894038807761553e-05,
|
| 331 |
+
"loss": 0.7614,
|
| 332 |
+
"step": 5400
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 0.11,
|
| 336 |
+
"learning_rate": 9.892038407681536e-05,
|
| 337 |
+
"loss": 0.8171,
|
| 338 |
+
"step": 5500
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.11,
|
| 342 |
+
"learning_rate": 9.890038007601521e-05,
|
| 343 |
+
"loss": 0.7649,
|
| 344 |
+
"step": 5600
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 0.11,
|
| 348 |
+
"learning_rate": 9.888037607521505e-05,
|
| 349 |
+
"loss": 0.8074,
|
| 350 |
+
"step": 5700
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 0.12,
|
| 354 |
+
"learning_rate": 9.886037207441488e-05,
|
| 355 |
+
"loss": 0.7591,
|
| 356 |
+
"step": 5800
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 0.12,
|
| 360 |
+
"learning_rate": 9.884036807361473e-05,
|
| 361 |
+
"loss": 0.8102,
|
| 362 |
+
"step": 5900
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 0.12,
|
| 366 |
+
"learning_rate": 9.882036407281456e-05,
|
| 367 |
+
"loss": 0.784,
|
| 368 |
+
"step": 6000
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 0.12,
|
| 372 |
+
"learning_rate": 9.88003600720144e-05,
|
| 373 |
+
"loss": 0.8367,
|
| 374 |
+
"step": 6100
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 0.12,
|
| 378 |
+
"learning_rate": 9.878035607121425e-05,
|
| 379 |
+
"loss": 0.8418,
|
| 380 |
+
"step": 6200
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.13,
|
| 384 |
+
"learning_rate": 9.876035207041409e-05,
|
| 385 |
+
"loss": 0.798,
|
| 386 |
+
"step": 6300
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 0.13,
|
| 390 |
+
"learning_rate": 9.874034806961394e-05,
|
| 391 |
+
"loss": 0.7928,
|
| 392 |
+
"step": 6400
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 0.13,
|
| 396 |
+
"learning_rate": 9.872034406881377e-05,
|
| 397 |
+
"loss": 0.8262,
|
| 398 |
+
"step": 6500
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 0.13,
|
| 402 |
+
"learning_rate": 9.870034006801361e-05,
|
| 403 |
+
"loss": 0.7965,
|
| 404 |
+
"step": 6600
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 0.13,
|
| 408 |
+
"learning_rate": 9.868033606721346e-05,
|
| 409 |
+
"loss": 0.7952,
|
| 410 |
+
"step": 6700
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 0.14,
|
| 414 |
+
"learning_rate": 9.866033206641329e-05,
|
| 415 |
+
"loss": 0.7987,
|
| 416 |
+
"step": 6800
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 0.14,
|
| 420 |
+
"learning_rate": 9.864032806561313e-05,
|
| 421 |
+
"loss": 0.7939,
|
| 422 |
+
"step": 6900
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.14,
|
| 426 |
+
"learning_rate": 9.862032406481296e-05,
|
| 427 |
+
"loss": 0.7484,
|
| 428 |
+
"step": 7000
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 0.14,
|
| 432 |
+
"learning_rate": 9.860032006401281e-05,
|
| 433 |
+
"loss": 0.819,
|
| 434 |
+
"step": 7100
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 0.14,
|
| 438 |
+
"learning_rate": 9.858031606321265e-05,
|
| 439 |
+
"loss": 0.747,
|
| 440 |
+
"step": 7200
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 0.15,
|
| 444 |
+
"learning_rate": 9.856031206241248e-05,
|
| 445 |
+
"loss": 0.8029,
|
| 446 |
+
"step": 7300
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 0.15,
|
| 450 |
+
"learning_rate": 9.854030806161233e-05,
|
| 451 |
+
"loss": 0.9013,
|
| 452 |
+
"step": 7400
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 0.15,
|
| 456 |
+
"learning_rate": 9.852030406081216e-05,
|
| 457 |
+
"loss": 0.8337,
|
| 458 |
+
"step": 7500
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 0.15,
|
| 462 |
+
"learning_rate": 9.8500300060012e-05,
|
| 463 |
+
"loss": 0.7927,
|
| 464 |
+
"step": 7600
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.15,
|
| 468 |
+
"learning_rate": 9.848029605921185e-05,
|
| 469 |
+
"loss": 0.8725,
|
| 470 |
+
"step": 7700
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 0.16,
|
| 474 |
+
"learning_rate": 9.846029205841169e-05,
|
| 475 |
+
"loss": 0.7865,
|
| 476 |
+
"step": 7800
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 0.16,
|
| 480 |
+
"learning_rate": 9.844028805761154e-05,
|
| 481 |
+
"loss": 0.7726,
|
| 482 |
+
"step": 7900
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 0.16,
|
| 486 |
+
"learning_rate": 9.842048409681936e-05,
|
| 487 |
+
"loss": 0.7868,
|
| 488 |
+
"step": 8000
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 0.16,
|
| 492 |
+
"learning_rate": 9.840048009601921e-05,
|
| 493 |
+
"loss": 0.7918,
|
| 494 |
+
"step": 8100
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 0.16,
|
| 498 |
+
"learning_rate": 9.838047609521905e-05,
|
| 499 |
+
"loss": 0.7924,
|
| 500 |
+
"step": 8200
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 0.17,
|
| 504 |
+
"learning_rate": 9.836047209441888e-05,
|
| 505 |
+
"loss": 0.8036,
|
| 506 |
+
"step": 8300
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 0.17,
|
| 510 |
+
"learning_rate": 9.834046809361873e-05,
|
| 511 |
+
"loss": 0.7844,
|
| 512 |
+
"step": 8400
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 0.17,
|
| 516 |
+
"learning_rate": 9.832046409281857e-05,
|
| 517 |
+
"loss": 0.8124,
|
| 518 |
+
"step": 8500
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 0.17,
|
| 522 |
+
"learning_rate": 9.830046009201842e-05,
|
| 523 |
+
"loss": 0.837,
|
| 524 |
+
"step": 8600
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 0.17,
|
| 528 |
+
"learning_rate": 9.828045609121826e-05,
|
| 529 |
+
"loss": 0.8315,
|
| 530 |
+
"step": 8700
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 0.18,
|
| 534 |
+
"learning_rate": 9.826045209041809e-05,
|
| 535 |
+
"loss": 0.7738,
|
| 536 |
+
"step": 8800
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 0.18,
|
| 540 |
+
"learning_rate": 9.824044808961794e-05,
|
| 541 |
+
"loss": 0.8164,
|
| 542 |
+
"step": 8900
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 0.18,
|
| 546 |
+
"learning_rate": 9.822044408881777e-05,
|
| 547 |
+
"loss": 0.7529,
|
| 548 |
+
"step": 9000
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 0.18,
|
| 552 |
+
"learning_rate": 9.820044008801761e-05,
|
| 553 |
+
"loss": 0.8037,
|
| 554 |
+
"step": 9100
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 0.18,
|
| 558 |
+
"learning_rate": 9.818043608721746e-05,
|
| 559 |
+
"loss": 0.8537,
|
| 560 |
+
"step": 9200
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 0.19,
|
| 564 |
+
"learning_rate": 9.816043208641729e-05,
|
| 565 |
+
"loss": 0.8241,
|
| 566 |
+
"step": 9300
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 0.19,
|
| 570 |
+
"learning_rate": 9.814042808561713e-05,
|
| 571 |
+
"loss": 0.7729,
|
| 572 |
+
"step": 9400
|
| 573 |
+
},
|
| 574 |
+
{
|
| 575 |
+
"epoch": 0.19,
|
| 576 |
+
"learning_rate": 9.812042408481696e-05,
|
| 577 |
+
"loss": 0.8116,
|
| 578 |
+
"step": 9500
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 0.19,
|
| 582 |
+
"learning_rate": 9.81004200840168e-05,
|
| 583 |
+
"loss": 0.8254,
|
| 584 |
+
"step": 9600
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 0.19,
|
| 588 |
+
"learning_rate": 9.808041608321665e-05,
|
| 589 |
+
"loss": 0.8285,
|
| 590 |
+
"step": 9700
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 0.2,
|
| 594 |
+
"learning_rate": 9.806041208241648e-05,
|
| 595 |
+
"loss": 0.7577,
|
| 596 |
+
"step": 9800
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 0.2,
|
| 600 |
+
"learning_rate": 9.804040808161633e-05,
|
| 601 |
+
"loss": 0.8024,
|
| 602 |
+
"step": 9900
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"epoch": 0.2,
|
| 606 |
+
"learning_rate": 9.802040408081616e-05,
|
| 607 |
+
"loss": 0.8078,
|
| 608 |
+
"step": 10000
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 0.2,
|
| 612 |
+
"eval_loss": 0.740916907787323,
|
| 613 |
+
"eval_runtime": 689.7578,
|
| 614 |
+
"eval_samples_per_second": 7.211,
|
| 615 |
+
"eval_steps_per_second": 0.902,
|
| 616 |
+
"step": 10000
|
| 617 |
+
}
|
| 618 |
+
],
|
| 619 |
+
"max_steps": 500000,
|
| 620 |
+
"num_train_epochs": 10,
|
| 621 |
+
"total_flos": 1.4097962687142298e+17,
|
| 622 |
+
"trial_name": null,
|
| 623 |
+
"trial_params": null
|
| 624 |
+
}
|
checkpoint-10000/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:86e3e950534e1e3e08a50aef6d282bea64ab70d530ec27d5c1bcac285ba164ae
|
| 3 |
+
size 4408
|
checkpoint-20000/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<pad>": 32000
|
| 3 |
+
}
|
checkpoint-20000/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:65a4f532f5ad3e6c930821b70e89a9df9eb72c675a2110100f9bf6a2ef582a09
|
| 3 |
+
size 134433530
|
checkpoint-20000/pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:de4c00e90d7499a8eb868bc357dfa0ad69fa679c6d4b245dbfd444739fe4ca38
|
| 3 |
+
size 67201802
|
checkpoint-20000/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:81e9bb7ffc314e87a0ca906b07b75a9b4b63be9b901ef0e9cb7e9f9cad382e03
|
| 3 |
+
size 14244
|
checkpoint-20000/scaler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c2a4165ea4d065ede7d3904be62b99a9319a89327b97a6f33a1b7258013f6996
|
| 3 |
+
size 988
|
checkpoint-20000/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:41efc1ee8f7840f17c8f469a1008689e2176bc8aef0637e7a3d176304878adf2
|
| 3 |
+
size 1064
|
checkpoint-20000/special_tokens_map.json
ADDED
|
@@ -0,0 +1,24 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"content": "<s>",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": false,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"eos_token": {
|
| 10 |
+
"content": "</s>",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"pad_token": "<s>",
|
| 17 |
+
"unk_token": {
|
| 18 |
+
"content": "<unk>",
|
| 19 |
+
"lstrip": false,
|
| 20 |
+
"normalized": false,
|
| 21 |
+
"rstrip": false,
|
| 22 |
+
"single_word": false
|
| 23 |
+
}
|
| 24 |
+
}
|
checkpoint-20000/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
checkpoint-20000/tokenizer.model
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9e556afd44213b6bd1be2b850ebbbd98f5481437a8021afaf58ee7fb1818d347
|
| 3 |
+
size 499723
|
checkpoint-20000/tokenizer_config.json
ADDED
|
@@ -0,0 +1,34 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_bos_token": true,
|
| 3 |
+
"add_eos_token": false,
|
| 4 |
+
"bos_token": {
|
| 5 |
+
"__type": "AddedToken",
|
| 6 |
+
"content": "<s>",
|
| 7 |
+
"lstrip": false,
|
| 8 |
+
"normalized": false,
|
| 9 |
+
"rstrip": false,
|
| 10 |
+
"single_word": false
|
| 11 |
+
},
|
| 12 |
+
"clean_up_tokenization_spaces": false,
|
| 13 |
+
"eos_token": {
|
| 14 |
+
"__type": "AddedToken",
|
| 15 |
+
"content": "</s>",
|
| 16 |
+
"lstrip": false,
|
| 17 |
+
"normalized": false,
|
| 18 |
+
"rstrip": false,
|
| 19 |
+
"single_word": false
|
| 20 |
+
},
|
| 21 |
+
"legacy": false,
|
| 22 |
+
"model_max_length": 1000000000000000019884624838656,
|
| 23 |
+
"pad_token": null,
|
| 24 |
+
"sp_model_kwargs": {},
|
| 25 |
+
"tokenizer_class": "LlamaTokenizer",
|
| 26 |
+
"unk_token": {
|
| 27 |
+
"__type": "AddedToken",
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false
|
| 33 |
+
}
|
| 34 |
+
}
|
checkpoint-20000/trainer_state.json
ADDED
|
@@ -0,0 +1,1232 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.7286937832832336,
|
| 3 |
+
"best_model_checkpoint": "/media/xzuyn/c41d93ed-83a7-4577-9d38-5062b7e14438/xzuyn/Documents/AI_ROCM_5-6/LLaMa-LoRA-Tuner/data/lora_models/LLaMa-2-WizardLM-Uncensored-Tulu-Format-50K-7B-LoRA/checkpoint-20000",
|
| 4 |
+
"epoch": 0.4,
|
| 5 |
+
"global_step": 20000,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.0,
|
| 12 |
+
"learning_rate": 0.0001,
|
| 13 |
+
"loss": 1.1136,
|
| 14 |
+
"step": 100
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.0,
|
| 18 |
+
"learning_rate": 9.997999599919985e-05,
|
| 19 |
+
"loss": 0.9595,
|
| 20 |
+
"step": 200
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.01,
|
| 24 |
+
"learning_rate": 9.995999199839968e-05,
|
| 25 |
+
"loss": 0.8936,
|
| 26 |
+
"step": 300
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.01,
|
| 30 |
+
"learning_rate": 9.993998799759952e-05,
|
| 31 |
+
"loss": 0.9237,
|
| 32 |
+
"step": 400
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.01,
|
| 36 |
+
"learning_rate": 9.992038407681536e-05,
|
| 37 |
+
"loss": 0.9005,
|
| 38 |
+
"step": 500
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.01,
|
| 42 |
+
"learning_rate": 9.99003800760152e-05,
|
| 43 |
+
"loss": 0.8433,
|
| 44 |
+
"step": 600
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.01,
|
| 48 |
+
"learning_rate": 9.988037607521505e-05,
|
| 49 |
+
"loss": 0.863,
|
| 50 |
+
"step": 700
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.02,
|
| 54 |
+
"learning_rate": 9.98603720744149e-05,
|
| 55 |
+
"loss": 0.8634,
|
| 56 |
+
"step": 800
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.02,
|
| 60 |
+
"learning_rate": 9.984036807361474e-05,
|
| 61 |
+
"loss": 0.8463,
|
| 62 |
+
"step": 900
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.02,
|
| 66 |
+
"learning_rate": 9.982036407281457e-05,
|
| 67 |
+
"loss": 0.8527,
|
| 68 |
+
"step": 1000
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.02,
|
| 72 |
+
"learning_rate": 9.980036007201441e-05,
|
| 73 |
+
"loss": 0.8259,
|
| 74 |
+
"step": 1100
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.02,
|
| 78 |
+
"learning_rate": 9.978035607121424e-05,
|
| 79 |
+
"loss": 0.8473,
|
| 80 |
+
"step": 1200
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.03,
|
| 84 |
+
"learning_rate": 9.976035207041409e-05,
|
| 85 |
+
"loss": 0.8462,
|
| 86 |
+
"step": 1300
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.03,
|
| 90 |
+
"learning_rate": 9.974034806961393e-05,
|
| 91 |
+
"loss": 0.8609,
|
| 92 |
+
"step": 1400
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.03,
|
| 96 |
+
"learning_rate": 9.972034406881376e-05,
|
| 97 |
+
"loss": 0.8238,
|
| 98 |
+
"step": 1500
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.03,
|
| 102 |
+
"learning_rate": 9.970034006801361e-05,
|
| 103 |
+
"loss": 0.8212,
|
| 104 |
+
"step": 1600
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 0.03,
|
| 108 |
+
"learning_rate": 9.968033606721344e-05,
|
| 109 |
+
"loss": 0.8391,
|
| 110 |
+
"step": 1700
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.04,
|
| 114 |
+
"learning_rate": 9.966033206641328e-05,
|
| 115 |
+
"loss": 0.774,
|
| 116 |
+
"step": 1800
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 0.04,
|
| 120 |
+
"learning_rate": 9.964032806561313e-05,
|
| 121 |
+
"loss": 0.8149,
|
| 122 |
+
"step": 1900
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.04,
|
| 126 |
+
"learning_rate": 9.962032406481296e-05,
|
| 127 |
+
"loss": 0.8794,
|
| 128 |
+
"step": 2000
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.04,
|
| 132 |
+
"learning_rate": 9.96003200640128e-05,
|
| 133 |
+
"loss": 0.7747,
|
| 134 |
+
"step": 2100
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.04,
|
| 138 |
+
"learning_rate": 9.958031606321264e-05,
|
| 139 |
+
"loss": 0.8191,
|
| 140 |
+
"step": 2200
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 0.05,
|
| 144 |
+
"learning_rate": 9.956031206241248e-05,
|
| 145 |
+
"loss": 0.8219,
|
| 146 |
+
"step": 2300
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 0.05,
|
| 150 |
+
"learning_rate": 9.954030806161232e-05,
|
| 151 |
+
"loss": 0.8191,
|
| 152 |
+
"step": 2400
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 0.05,
|
| 156 |
+
"learning_rate": 9.952030406081217e-05,
|
| 157 |
+
"loss": 0.8662,
|
| 158 |
+
"step": 2500
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 0.05,
|
| 162 |
+
"learning_rate": 9.950030006001201e-05,
|
| 163 |
+
"loss": 0.787,
|
| 164 |
+
"step": 2600
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.05,
|
| 168 |
+
"learning_rate": 9.948029605921184e-05,
|
| 169 |
+
"loss": 0.8668,
|
| 170 |
+
"step": 2700
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.06,
|
| 174 |
+
"learning_rate": 9.946029205841169e-05,
|
| 175 |
+
"loss": 0.7817,
|
| 176 |
+
"step": 2800
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 0.06,
|
| 180 |
+
"learning_rate": 9.944028805761153e-05,
|
| 181 |
+
"loss": 0.8021,
|
| 182 |
+
"step": 2900
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 0.06,
|
| 186 |
+
"learning_rate": 9.942028405681136e-05,
|
| 187 |
+
"loss": 0.8324,
|
| 188 |
+
"step": 3000
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 0.06,
|
| 192 |
+
"learning_rate": 9.940028005601121e-05,
|
| 193 |
+
"loss": 0.8631,
|
| 194 |
+
"step": 3100
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 0.06,
|
| 198 |
+
"learning_rate": 9.938027605521105e-05,
|
| 199 |
+
"loss": 0.797,
|
| 200 |
+
"step": 3200
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 0.07,
|
| 204 |
+
"learning_rate": 9.936027205441088e-05,
|
| 205 |
+
"loss": 0.8224,
|
| 206 |
+
"step": 3300
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.07,
|
| 210 |
+
"learning_rate": 9.934026805361073e-05,
|
| 211 |
+
"loss": 0.8164,
|
| 212 |
+
"step": 3400
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.07,
|
| 216 |
+
"learning_rate": 9.932026405281056e-05,
|
| 217 |
+
"loss": 0.7975,
|
| 218 |
+
"step": 3500
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 0.07,
|
| 222 |
+
"learning_rate": 9.930046009201841e-05,
|
| 223 |
+
"loss": 0.8468,
|
| 224 |
+
"step": 3600
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 0.07,
|
| 228 |
+
"learning_rate": 9.928045609121824e-05,
|
| 229 |
+
"loss": 0.8169,
|
| 230 |
+
"step": 3700
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 0.08,
|
| 234 |
+
"learning_rate": 9.926045209041809e-05,
|
| 235 |
+
"loss": 0.7761,
|
| 236 |
+
"step": 3800
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 0.08,
|
| 240 |
+
"learning_rate": 9.924044808961793e-05,
|
| 241 |
+
"loss": 0.8042,
|
| 242 |
+
"step": 3900
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 0.08,
|
| 246 |
+
"learning_rate": 9.922044408881776e-05,
|
| 247 |
+
"loss": 0.8146,
|
| 248 |
+
"step": 4000
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 0.08,
|
| 252 |
+
"learning_rate": 9.920044008801761e-05,
|
| 253 |
+
"loss": 0.8243,
|
| 254 |
+
"step": 4100
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.08,
|
| 258 |
+
"learning_rate": 9.918043608721744e-05,
|
| 259 |
+
"loss": 0.8315,
|
| 260 |
+
"step": 4200
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 0.09,
|
| 264 |
+
"learning_rate": 9.916043208641728e-05,
|
| 265 |
+
"loss": 0.8014,
|
| 266 |
+
"step": 4300
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 0.09,
|
| 270 |
+
"learning_rate": 9.914042808561713e-05,
|
| 271 |
+
"loss": 0.8241,
|
| 272 |
+
"step": 4400
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 0.09,
|
| 276 |
+
"learning_rate": 9.912042408481696e-05,
|
| 277 |
+
"loss": 0.8233,
|
| 278 |
+
"step": 4500
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 0.09,
|
| 282 |
+
"learning_rate": 9.91004200840168e-05,
|
| 283 |
+
"loss": 0.7748,
|
| 284 |
+
"step": 4600
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 0.09,
|
| 288 |
+
"learning_rate": 9.908041608321665e-05,
|
| 289 |
+
"loss": 0.8197,
|
| 290 |
+
"step": 4700
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 0.1,
|
| 294 |
+
"learning_rate": 9.906041208241649e-05,
|
| 295 |
+
"loss": 0.8195,
|
| 296 |
+
"step": 4800
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.1,
|
| 300 |
+
"learning_rate": 9.904040808161634e-05,
|
| 301 |
+
"loss": 0.7864,
|
| 302 |
+
"step": 4900
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 0.1,
|
| 306 |
+
"learning_rate": 9.902040408081617e-05,
|
| 307 |
+
"loss": 0.7812,
|
| 308 |
+
"step": 5000
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 0.1,
|
| 312 |
+
"learning_rate": 9.900040008001601e-05,
|
| 313 |
+
"loss": 0.7966,
|
| 314 |
+
"step": 5100
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 0.1,
|
| 318 |
+
"learning_rate": 9.898039607921584e-05,
|
| 319 |
+
"loss": 0.793,
|
| 320 |
+
"step": 5200
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 0.11,
|
| 324 |
+
"learning_rate": 9.896039207841569e-05,
|
| 325 |
+
"loss": 0.8568,
|
| 326 |
+
"step": 5300
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 0.11,
|
| 330 |
+
"learning_rate": 9.894038807761553e-05,
|
| 331 |
+
"loss": 0.7614,
|
| 332 |
+
"step": 5400
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 0.11,
|
| 336 |
+
"learning_rate": 9.892038407681536e-05,
|
| 337 |
+
"loss": 0.8171,
|
| 338 |
+
"step": 5500
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.11,
|
| 342 |
+
"learning_rate": 9.890038007601521e-05,
|
| 343 |
+
"loss": 0.7649,
|
| 344 |
+
"step": 5600
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 0.11,
|
| 348 |
+
"learning_rate": 9.888037607521505e-05,
|
| 349 |
+
"loss": 0.8074,
|
| 350 |
+
"step": 5700
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 0.12,
|
| 354 |
+
"learning_rate": 9.886037207441488e-05,
|
| 355 |
+
"loss": 0.7591,
|
| 356 |
+
"step": 5800
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 0.12,
|
| 360 |
+
"learning_rate": 9.884036807361473e-05,
|
| 361 |
+
"loss": 0.8102,
|
| 362 |
+
"step": 5900
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 0.12,
|
| 366 |
+
"learning_rate": 9.882036407281456e-05,
|
| 367 |
+
"loss": 0.784,
|
| 368 |
+
"step": 6000
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 0.12,
|
| 372 |
+
"learning_rate": 9.88003600720144e-05,
|
| 373 |
+
"loss": 0.8367,
|
| 374 |
+
"step": 6100
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 0.12,
|
| 378 |
+
"learning_rate": 9.878035607121425e-05,
|
| 379 |
+
"loss": 0.8418,
|
| 380 |
+
"step": 6200
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.13,
|
| 384 |
+
"learning_rate": 9.876035207041409e-05,
|
| 385 |
+
"loss": 0.798,
|
| 386 |
+
"step": 6300
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 0.13,
|
| 390 |
+
"learning_rate": 9.874034806961394e-05,
|
| 391 |
+
"loss": 0.7928,
|
| 392 |
+
"step": 6400
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 0.13,
|
| 396 |
+
"learning_rate": 9.872034406881377e-05,
|
| 397 |
+
"loss": 0.8262,
|
| 398 |
+
"step": 6500
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 0.13,
|
| 402 |
+
"learning_rate": 9.870034006801361e-05,
|
| 403 |
+
"loss": 0.7965,
|
| 404 |
+
"step": 6600
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 0.13,
|
| 408 |
+
"learning_rate": 9.868033606721346e-05,
|
| 409 |
+
"loss": 0.7952,
|
| 410 |
+
"step": 6700
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 0.14,
|
| 414 |
+
"learning_rate": 9.866033206641329e-05,
|
| 415 |
+
"loss": 0.7987,
|
| 416 |
+
"step": 6800
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 0.14,
|
| 420 |
+
"learning_rate": 9.864032806561313e-05,
|
| 421 |
+
"loss": 0.7939,
|
| 422 |
+
"step": 6900
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.14,
|
| 426 |
+
"learning_rate": 9.862032406481296e-05,
|
| 427 |
+
"loss": 0.7484,
|
| 428 |
+
"step": 7000
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 0.14,
|
| 432 |
+
"learning_rate": 9.860032006401281e-05,
|
| 433 |
+
"loss": 0.819,
|
| 434 |
+
"step": 7100
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 0.14,
|
| 438 |
+
"learning_rate": 9.858031606321265e-05,
|
| 439 |
+
"loss": 0.747,
|
| 440 |
+
"step": 7200
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 0.15,
|
| 444 |
+
"learning_rate": 9.856031206241248e-05,
|
| 445 |
+
"loss": 0.8029,
|
| 446 |
+
"step": 7300
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 0.15,
|
| 450 |
+
"learning_rate": 9.854030806161233e-05,
|
| 451 |
+
"loss": 0.9013,
|
| 452 |
+
"step": 7400
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 0.15,
|
| 456 |
+
"learning_rate": 9.852030406081216e-05,
|
| 457 |
+
"loss": 0.8337,
|
| 458 |
+
"step": 7500
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 0.15,
|
| 462 |
+
"learning_rate": 9.8500300060012e-05,
|
| 463 |
+
"loss": 0.7927,
|
| 464 |
+
"step": 7600
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.15,
|
| 468 |
+
"learning_rate": 9.848029605921185e-05,
|
| 469 |
+
"loss": 0.8725,
|
| 470 |
+
"step": 7700
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 0.16,
|
| 474 |
+
"learning_rate": 9.846029205841169e-05,
|
| 475 |
+
"loss": 0.7865,
|
| 476 |
+
"step": 7800
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 0.16,
|
| 480 |
+
"learning_rate": 9.844028805761154e-05,
|
| 481 |
+
"loss": 0.7726,
|
| 482 |
+
"step": 7900
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 0.16,
|
| 486 |
+
"learning_rate": 9.842048409681936e-05,
|
| 487 |
+
"loss": 0.7868,
|
| 488 |
+
"step": 8000
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 0.16,
|
| 492 |
+
"learning_rate": 9.840048009601921e-05,
|
| 493 |
+
"loss": 0.7918,
|
| 494 |
+
"step": 8100
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 0.16,
|
| 498 |
+
"learning_rate": 9.838047609521905e-05,
|
| 499 |
+
"loss": 0.7924,
|
| 500 |
+
"step": 8200
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 0.17,
|
| 504 |
+
"learning_rate": 9.836047209441888e-05,
|
| 505 |
+
"loss": 0.8036,
|
| 506 |
+
"step": 8300
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 0.17,
|
| 510 |
+
"learning_rate": 9.834046809361873e-05,
|
| 511 |
+
"loss": 0.7844,
|
| 512 |
+
"step": 8400
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 0.17,
|
| 516 |
+
"learning_rate": 9.832046409281857e-05,
|
| 517 |
+
"loss": 0.8124,
|
| 518 |
+
"step": 8500
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 0.17,
|
| 522 |
+
"learning_rate": 9.830046009201842e-05,
|
| 523 |
+
"loss": 0.837,
|
| 524 |
+
"step": 8600
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 0.17,
|
| 528 |
+
"learning_rate": 9.828045609121826e-05,
|
| 529 |
+
"loss": 0.8315,
|
| 530 |
+
"step": 8700
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 0.18,
|
| 534 |
+
"learning_rate": 9.826045209041809e-05,
|
| 535 |
+
"loss": 0.7738,
|
| 536 |
+
"step": 8800
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 0.18,
|
| 540 |
+
"learning_rate": 9.824044808961794e-05,
|
| 541 |
+
"loss": 0.8164,
|
| 542 |
+
"step": 8900
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 0.18,
|
| 546 |
+
"learning_rate": 9.822044408881777e-05,
|
| 547 |
+
"loss": 0.7529,
|
| 548 |
+
"step": 9000
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 0.18,
|
| 552 |
+
"learning_rate": 9.820044008801761e-05,
|
| 553 |
+
"loss": 0.8037,
|
| 554 |
+
"step": 9100
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 0.18,
|
| 558 |
+
"learning_rate": 9.818043608721746e-05,
|
| 559 |
+
"loss": 0.8537,
|
| 560 |
+
"step": 9200
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 0.19,
|
| 564 |
+
"learning_rate": 9.816043208641729e-05,
|
| 565 |
+
"loss": 0.8241,
|
| 566 |
+
"step": 9300
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 0.19,
|
| 570 |
+
"learning_rate": 9.814042808561713e-05,
|
| 571 |
+
"loss": 0.7729,
|
| 572 |
+
"step": 9400
|
| 573 |
+
},
|
| 574 |
+
{
|
| 575 |
+
"epoch": 0.19,
|
| 576 |
+
"learning_rate": 9.812042408481696e-05,
|
| 577 |
+
"loss": 0.8116,
|
| 578 |
+
"step": 9500
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 0.19,
|
| 582 |
+
"learning_rate": 9.81004200840168e-05,
|
| 583 |
+
"loss": 0.8254,
|
| 584 |
+
"step": 9600
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 0.19,
|
| 588 |
+
"learning_rate": 9.808041608321665e-05,
|
| 589 |
+
"loss": 0.8285,
|
| 590 |
+
"step": 9700
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 0.2,
|
| 594 |
+
"learning_rate": 9.806041208241648e-05,
|
| 595 |
+
"loss": 0.7577,
|
| 596 |
+
"step": 9800
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 0.2,
|
| 600 |
+
"learning_rate": 9.804040808161633e-05,
|
| 601 |
+
"loss": 0.8024,
|
| 602 |
+
"step": 9900
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"epoch": 0.2,
|
| 606 |
+
"learning_rate": 9.802040408081616e-05,
|
| 607 |
+
"loss": 0.8078,
|
| 608 |
+
"step": 10000
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 0.2,
|
| 612 |
+
"eval_loss": 0.740916907787323,
|
| 613 |
+
"eval_runtime": 689.7578,
|
| 614 |
+
"eval_samples_per_second": 7.211,
|
| 615 |
+
"eval_steps_per_second": 0.902,
|
| 616 |
+
"step": 10000
|
| 617 |
+
},
|
| 618 |
+
{
|
| 619 |
+
"epoch": 0.2,
|
| 620 |
+
"learning_rate": 9.8000400080016e-05,
|
| 621 |
+
"loss": 0.7849,
|
| 622 |
+
"step": 10100
|
| 623 |
+
},
|
| 624 |
+
{
|
| 625 |
+
"epoch": 0.2,
|
| 626 |
+
"learning_rate": 9.798039607921585e-05,
|
| 627 |
+
"loss": 0.7931,
|
| 628 |
+
"step": 10200
|
| 629 |
+
},
|
| 630 |
+
{
|
| 631 |
+
"epoch": 0.21,
|
| 632 |
+
"learning_rate": 9.796039207841569e-05,
|
| 633 |
+
"loss": 0.7884,
|
| 634 |
+
"step": 10300
|
| 635 |
+
},
|
| 636 |
+
{
|
| 637 |
+
"epoch": 0.21,
|
| 638 |
+
"learning_rate": 9.794038807761554e-05,
|
| 639 |
+
"loss": 0.7949,
|
| 640 |
+
"step": 10400
|
| 641 |
+
},
|
| 642 |
+
{
|
| 643 |
+
"epoch": 0.21,
|
| 644 |
+
"learning_rate": 9.792038407681537e-05,
|
| 645 |
+
"loss": 0.7916,
|
| 646 |
+
"step": 10500
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"epoch": 0.21,
|
| 650 |
+
"learning_rate": 9.790038007601521e-05,
|
| 651 |
+
"loss": 0.8072,
|
| 652 |
+
"step": 10600
|
| 653 |
+
},
|
| 654 |
+
{
|
| 655 |
+
"epoch": 0.21,
|
| 656 |
+
"learning_rate": 9.788037607521506e-05,
|
| 657 |
+
"loss": 0.7965,
|
| 658 |
+
"step": 10700
|
| 659 |
+
},
|
| 660 |
+
{
|
| 661 |
+
"epoch": 0.22,
|
| 662 |
+
"learning_rate": 9.786037207441489e-05,
|
| 663 |
+
"loss": 0.8039,
|
| 664 |
+
"step": 10800
|
| 665 |
+
},
|
| 666 |
+
{
|
| 667 |
+
"epoch": 0.22,
|
| 668 |
+
"learning_rate": 9.784036807361473e-05,
|
| 669 |
+
"loss": 0.8011,
|
| 670 |
+
"step": 10900
|
| 671 |
+
},
|
| 672 |
+
{
|
| 673 |
+
"epoch": 0.22,
|
| 674 |
+
"learning_rate": 9.782036407281456e-05,
|
| 675 |
+
"loss": 0.8014,
|
| 676 |
+
"step": 11000
|
| 677 |
+
},
|
| 678 |
+
{
|
| 679 |
+
"epoch": 0.22,
|
| 680 |
+
"learning_rate": 9.78003600720144e-05,
|
| 681 |
+
"loss": 0.7835,
|
| 682 |
+
"step": 11100
|
| 683 |
+
},
|
| 684 |
+
{
|
| 685 |
+
"epoch": 0.22,
|
| 686 |
+
"learning_rate": 9.778035607121425e-05,
|
| 687 |
+
"loss": 0.8469,
|
| 688 |
+
"step": 11200
|
| 689 |
+
},
|
| 690 |
+
{
|
| 691 |
+
"epoch": 0.23,
|
| 692 |
+
"learning_rate": 9.776035207041408e-05,
|
| 693 |
+
"loss": 0.8018,
|
| 694 |
+
"step": 11300
|
| 695 |
+
},
|
| 696 |
+
{
|
| 697 |
+
"epoch": 0.23,
|
| 698 |
+
"learning_rate": 9.774034806961393e-05,
|
| 699 |
+
"loss": 0.7839,
|
| 700 |
+
"step": 11400
|
| 701 |
+
},
|
| 702 |
+
{
|
| 703 |
+
"epoch": 0.23,
|
| 704 |
+
"learning_rate": 9.772034406881376e-05,
|
| 705 |
+
"loss": 0.8312,
|
| 706 |
+
"step": 11500
|
| 707 |
+
},
|
| 708 |
+
{
|
| 709 |
+
"epoch": 0.23,
|
| 710 |
+
"learning_rate": 9.770054010802161e-05,
|
| 711 |
+
"loss": 0.7886,
|
| 712 |
+
"step": 11600
|
| 713 |
+
},
|
| 714 |
+
{
|
| 715 |
+
"epoch": 0.23,
|
| 716 |
+
"learning_rate": 9.768053610722146e-05,
|
| 717 |
+
"loss": 0.8096,
|
| 718 |
+
"step": 11700
|
| 719 |
+
},
|
| 720 |
+
{
|
| 721 |
+
"epoch": 0.24,
|
| 722 |
+
"learning_rate": 9.766053210642129e-05,
|
| 723 |
+
"loss": 0.7806,
|
| 724 |
+
"step": 11800
|
| 725 |
+
},
|
| 726 |
+
{
|
| 727 |
+
"epoch": 0.24,
|
| 728 |
+
"learning_rate": 9.764052810562113e-05,
|
| 729 |
+
"loss": 0.8512,
|
| 730 |
+
"step": 11900
|
| 731 |
+
},
|
| 732 |
+
{
|
| 733 |
+
"epoch": 0.24,
|
| 734 |
+
"learning_rate": 9.762052410482096e-05,
|
| 735 |
+
"loss": 0.8475,
|
| 736 |
+
"step": 12000
|
| 737 |
+
},
|
| 738 |
+
{
|
| 739 |
+
"epoch": 0.24,
|
| 740 |
+
"learning_rate": 9.76005201040208e-05,
|
| 741 |
+
"loss": 0.749,
|
| 742 |
+
"step": 12100
|
| 743 |
+
},
|
| 744 |
+
{
|
| 745 |
+
"epoch": 0.24,
|
| 746 |
+
"learning_rate": 9.758051610322065e-05,
|
| 747 |
+
"loss": 0.8245,
|
| 748 |
+
"step": 12200
|
| 749 |
+
},
|
| 750 |
+
{
|
| 751 |
+
"epoch": 0.25,
|
| 752 |
+
"learning_rate": 9.756051210242048e-05,
|
| 753 |
+
"loss": 0.7966,
|
| 754 |
+
"step": 12300
|
| 755 |
+
},
|
| 756 |
+
{
|
| 757 |
+
"epoch": 0.25,
|
| 758 |
+
"learning_rate": 9.754050810162033e-05,
|
| 759 |
+
"loss": 0.773,
|
| 760 |
+
"step": 12400
|
| 761 |
+
},
|
| 762 |
+
{
|
| 763 |
+
"epoch": 0.25,
|
| 764 |
+
"learning_rate": 9.752050410082017e-05,
|
| 765 |
+
"loss": 0.7876,
|
| 766 |
+
"step": 12500
|
| 767 |
+
},
|
| 768 |
+
{
|
| 769 |
+
"epoch": 0.25,
|
| 770 |
+
"learning_rate": 9.750050010002001e-05,
|
| 771 |
+
"loss": 0.7982,
|
| 772 |
+
"step": 12600
|
| 773 |
+
},
|
| 774 |
+
{
|
| 775 |
+
"epoch": 0.25,
|
| 776 |
+
"learning_rate": 9.748049609921986e-05,
|
| 777 |
+
"loss": 0.7509,
|
| 778 |
+
"step": 12700
|
| 779 |
+
},
|
| 780 |
+
{
|
| 781 |
+
"epoch": 0.26,
|
| 782 |
+
"learning_rate": 9.746049209841969e-05,
|
| 783 |
+
"loss": 0.7815,
|
| 784 |
+
"step": 12800
|
| 785 |
+
},
|
| 786 |
+
{
|
| 787 |
+
"epoch": 0.26,
|
| 788 |
+
"learning_rate": 9.744048809761953e-05,
|
| 789 |
+
"loss": 0.8516,
|
| 790 |
+
"step": 12900
|
| 791 |
+
},
|
| 792 |
+
{
|
| 793 |
+
"epoch": 0.26,
|
| 794 |
+
"learning_rate": 9.742048409681937e-05,
|
| 795 |
+
"loss": 0.7684,
|
| 796 |
+
"step": 13000
|
| 797 |
+
},
|
| 798 |
+
{
|
| 799 |
+
"epoch": 0.26,
|
| 800 |
+
"learning_rate": 9.740048009601921e-05,
|
| 801 |
+
"loss": 0.7889,
|
| 802 |
+
"step": 13100
|
| 803 |
+
},
|
| 804 |
+
{
|
| 805 |
+
"epoch": 0.26,
|
| 806 |
+
"learning_rate": 9.738047609521905e-05,
|
| 807 |
+
"loss": 0.7882,
|
| 808 |
+
"step": 13200
|
| 809 |
+
},
|
| 810 |
+
{
|
| 811 |
+
"epoch": 0.27,
|
| 812 |
+
"learning_rate": 9.736047209441889e-05,
|
| 813 |
+
"loss": 0.8196,
|
| 814 |
+
"step": 13300
|
| 815 |
+
},
|
| 816 |
+
{
|
| 817 |
+
"epoch": 0.27,
|
| 818 |
+
"learning_rate": 9.734046809361873e-05,
|
| 819 |
+
"loss": 0.7889,
|
| 820 |
+
"step": 13400
|
| 821 |
+
},
|
| 822 |
+
{
|
| 823 |
+
"epoch": 0.27,
|
| 824 |
+
"learning_rate": 9.732046409281856e-05,
|
| 825 |
+
"loss": 0.7931,
|
| 826 |
+
"step": 13500
|
| 827 |
+
},
|
| 828 |
+
{
|
| 829 |
+
"epoch": 0.27,
|
| 830 |
+
"learning_rate": 9.73004600920184e-05,
|
| 831 |
+
"loss": 0.8172,
|
| 832 |
+
"step": 13600
|
| 833 |
+
},
|
| 834 |
+
{
|
| 835 |
+
"epoch": 0.27,
|
| 836 |
+
"learning_rate": 9.728045609121825e-05,
|
| 837 |
+
"loss": 0.8364,
|
| 838 |
+
"step": 13700
|
| 839 |
+
},
|
| 840 |
+
{
|
| 841 |
+
"epoch": 0.28,
|
| 842 |
+
"learning_rate": 9.726045209041808e-05,
|
| 843 |
+
"loss": 0.822,
|
| 844 |
+
"step": 13800
|
| 845 |
+
},
|
| 846 |
+
{
|
| 847 |
+
"epoch": 0.28,
|
| 848 |
+
"learning_rate": 9.724044808961793e-05,
|
| 849 |
+
"loss": 0.8546,
|
| 850 |
+
"step": 13900
|
| 851 |
+
},
|
| 852 |
+
{
|
| 853 |
+
"epoch": 0.28,
|
| 854 |
+
"learning_rate": 9.722044408881777e-05,
|
| 855 |
+
"loss": 0.7977,
|
| 856 |
+
"step": 14000
|
| 857 |
+
},
|
| 858 |
+
{
|
| 859 |
+
"epoch": 0.28,
|
| 860 |
+
"learning_rate": 9.720044008801761e-05,
|
| 861 |
+
"loss": 0.7591,
|
| 862 |
+
"step": 14100
|
| 863 |
+
},
|
| 864 |
+
{
|
| 865 |
+
"epoch": 0.28,
|
| 866 |
+
"learning_rate": 9.718063612722545e-05,
|
| 867 |
+
"loss": 0.7787,
|
| 868 |
+
"step": 14200
|
| 869 |
+
},
|
| 870 |
+
{
|
| 871 |
+
"epoch": 0.29,
|
| 872 |
+
"learning_rate": 9.716063212642529e-05,
|
| 873 |
+
"loss": 0.8361,
|
| 874 |
+
"step": 14300
|
| 875 |
+
},
|
| 876 |
+
{
|
| 877 |
+
"epoch": 0.29,
|
| 878 |
+
"learning_rate": 9.714062812562513e-05,
|
| 879 |
+
"loss": 0.8153,
|
| 880 |
+
"step": 14400
|
| 881 |
+
},
|
| 882 |
+
{
|
| 883 |
+
"epoch": 0.29,
|
| 884 |
+
"learning_rate": 9.712062412482496e-05,
|
| 885 |
+
"loss": 0.776,
|
| 886 |
+
"step": 14500
|
| 887 |
+
},
|
| 888 |
+
{
|
| 889 |
+
"epoch": 0.29,
|
| 890 |
+
"learning_rate": 9.71006201240248e-05,
|
| 891 |
+
"loss": 0.7491,
|
| 892 |
+
"step": 14600
|
| 893 |
+
},
|
| 894 |
+
{
|
| 895 |
+
"epoch": 0.29,
|
| 896 |
+
"learning_rate": 9.708061612322465e-05,
|
| 897 |
+
"loss": 0.7975,
|
| 898 |
+
"step": 14700
|
| 899 |
+
},
|
| 900 |
+
{
|
| 901 |
+
"epoch": 0.3,
|
| 902 |
+
"learning_rate": 9.70606121224245e-05,
|
| 903 |
+
"loss": 0.7637,
|
| 904 |
+
"step": 14800
|
| 905 |
+
},
|
| 906 |
+
{
|
| 907 |
+
"epoch": 0.3,
|
| 908 |
+
"learning_rate": 9.704060812162434e-05,
|
| 909 |
+
"loss": 0.7977,
|
| 910 |
+
"step": 14900
|
| 911 |
+
},
|
| 912 |
+
{
|
| 913 |
+
"epoch": 0.3,
|
| 914 |
+
"learning_rate": 9.702060412082417e-05,
|
| 915 |
+
"loss": 0.8041,
|
| 916 |
+
"step": 15000
|
| 917 |
+
},
|
| 918 |
+
{
|
| 919 |
+
"epoch": 0.3,
|
| 920 |
+
"learning_rate": 9.700060012002401e-05,
|
| 921 |
+
"loss": 0.7997,
|
| 922 |
+
"step": 15100
|
| 923 |
+
},
|
| 924 |
+
{
|
| 925 |
+
"epoch": 0.3,
|
| 926 |
+
"learning_rate": 9.698059611922386e-05,
|
| 927 |
+
"loss": 0.8221,
|
| 928 |
+
"step": 15200
|
| 929 |
+
},
|
| 930 |
+
{
|
| 931 |
+
"epoch": 0.31,
|
| 932 |
+
"learning_rate": 9.696059211842369e-05,
|
| 933 |
+
"loss": 0.7657,
|
| 934 |
+
"step": 15300
|
| 935 |
+
},
|
| 936 |
+
{
|
| 937 |
+
"epoch": 0.31,
|
| 938 |
+
"learning_rate": 9.694058811762353e-05,
|
| 939 |
+
"loss": 0.8091,
|
| 940 |
+
"step": 15400
|
| 941 |
+
},
|
| 942 |
+
{
|
| 943 |
+
"epoch": 0.31,
|
| 944 |
+
"learning_rate": 9.692058411682336e-05,
|
| 945 |
+
"loss": 0.7692,
|
| 946 |
+
"step": 15500
|
| 947 |
+
},
|
| 948 |
+
{
|
| 949 |
+
"epoch": 0.31,
|
| 950 |
+
"learning_rate": 9.690058011602321e-05,
|
| 951 |
+
"loss": 0.7988,
|
| 952 |
+
"step": 15600
|
| 953 |
+
},
|
| 954 |
+
{
|
| 955 |
+
"epoch": 0.31,
|
| 956 |
+
"learning_rate": 9.688057611522305e-05,
|
| 957 |
+
"loss": 0.7457,
|
| 958 |
+
"step": 15700
|
| 959 |
+
},
|
| 960 |
+
{
|
| 961 |
+
"epoch": 0.32,
|
| 962 |
+
"learning_rate": 9.686057211442288e-05,
|
| 963 |
+
"loss": 0.8223,
|
| 964 |
+
"step": 15800
|
| 965 |
+
},
|
| 966 |
+
{
|
| 967 |
+
"epoch": 0.32,
|
| 968 |
+
"learning_rate": 9.684056811362273e-05,
|
| 969 |
+
"loss": 0.8366,
|
| 970 |
+
"step": 15900
|
| 971 |
+
},
|
| 972 |
+
{
|
| 973 |
+
"epoch": 0.32,
|
| 974 |
+
"learning_rate": 9.682056411282256e-05,
|
| 975 |
+
"loss": 0.7959,
|
| 976 |
+
"step": 16000
|
| 977 |
+
},
|
| 978 |
+
{
|
| 979 |
+
"epoch": 0.32,
|
| 980 |
+
"learning_rate": 9.680076015203041e-05,
|
| 981 |
+
"loss": 0.7944,
|
| 982 |
+
"step": 16100
|
| 983 |
+
},
|
| 984 |
+
{
|
| 985 |
+
"epoch": 0.32,
|
| 986 |
+
"learning_rate": 9.678075615123024e-05,
|
| 987 |
+
"loss": 0.7645,
|
| 988 |
+
"step": 16200
|
| 989 |
+
},
|
| 990 |
+
{
|
| 991 |
+
"epoch": 0.33,
|
| 992 |
+
"learning_rate": 9.676075215043009e-05,
|
| 993 |
+
"loss": 0.8111,
|
| 994 |
+
"step": 16300
|
| 995 |
+
},
|
| 996 |
+
{
|
| 997 |
+
"epoch": 0.33,
|
| 998 |
+
"learning_rate": 9.674074814962993e-05,
|
| 999 |
+
"loss": 0.7907,
|
| 1000 |
+
"step": 16400
|
| 1001 |
+
},
|
| 1002 |
+
{
|
| 1003 |
+
"epoch": 0.33,
|
| 1004 |
+
"learning_rate": 9.672074414882976e-05,
|
| 1005 |
+
"loss": 0.8108,
|
| 1006 |
+
"step": 16500
|
| 1007 |
+
},
|
| 1008 |
+
{
|
| 1009 |
+
"epoch": 0.33,
|
| 1010 |
+
"learning_rate": 9.670074014802961e-05,
|
| 1011 |
+
"loss": 0.8119,
|
| 1012 |
+
"step": 16600
|
| 1013 |
+
},
|
| 1014 |
+
{
|
| 1015 |
+
"epoch": 0.33,
|
| 1016 |
+
"learning_rate": 9.668073614722945e-05,
|
| 1017 |
+
"loss": 0.8109,
|
| 1018 |
+
"step": 16700
|
| 1019 |
+
},
|
| 1020 |
+
{
|
| 1021 |
+
"epoch": 0.34,
|
| 1022 |
+
"learning_rate": 9.666073214642928e-05,
|
| 1023 |
+
"loss": 0.7862,
|
| 1024 |
+
"step": 16800
|
| 1025 |
+
},
|
| 1026 |
+
{
|
| 1027 |
+
"epoch": 0.34,
|
| 1028 |
+
"learning_rate": 9.664072814562913e-05,
|
| 1029 |
+
"loss": 0.7886,
|
| 1030 |
+
"step": 16900
|
| 1031 |
+
},
|
| 1032 |
+
{
|
| 1033 |
+
"epoch": 0.34,
|
| 1034 |
+
"learning_rate": 9.662072414482897e-05,
|
| 1035 |
+
"loss": 0.786,
|
| 1036 |
+
"step": 17000
|
| 1037 |
+
},
|
| 1038 |
+
{
|
| 1039 |
+
"epoch": 0.34,
|
| 1040 |
+
"learning_rate": 9.660072014402882e-05,
|
| 1041 |
+
"loss": 0.8014,
|
| 1042 |
+
"step": 17100
|
| 1043 |
+
},
|
| 1044 |
+
{
|
| 1045 |
+
"epoch": 0.34,
|
| 1046 |
+
"learning_rate": 9.658071614322865e-05,
|
| 1047 |
+
"loss": 0.7486,
|
| 1048 |
+
"step": 17200
|
| 1049 |
+
},
|
| 1050 |
+
{
|
| 1051 |
+
"epoch": 0.35,
|
| 1052 |
+
"learning_rate": 9.65607121424285e-05,
|
| 1053 |
+
"loss": 0.7688,
|
| 1054 |
+
"step": 17300
|
| 1055 |
+
},
|
| 1056 |
+
{
|
| 1057 |
+
"epoch": 0.35,
|
| 1058 |
+
"learning_rate": 9.654070814162834e-05,
|
| 1059 |
+
"loss": 0.7815,
|
| 1060 |
+
"step": 17400
|
| 1061 |
+
},
|
| 1062 |
+
{
|
| 1063 |
+
"epoch": 0.35,
|
| 1064 |
+
"learning_rate": 9.652070414082817e-05,
|
| 1065 |
+
"loss": 0.7857,
|
| 1066 |
+
"step": 17500
|
| 1067 |
+
},
|
| 1068 |
+
{
|
| 1069 |
+
"epoch": 0.35,
|
| 1070 |
+
"learning_rate": 9.650070014002801e-05,
|
| 1071 |
+
"loss": 0.777,
|
| 1072 |
+
"step": 17600
|
| 1073 |
+
},
|
| 1074 |
+
{
|
| 1075 |
+
"epoch": 0.35,
|
| 1076 |
+
"learning_rate": 9.648069613922786e-05,
|
| 1077 |
+
"loss": 0.8186,
|
| 1078 |
+
"step": 17700
|
| 1079 |
+
},
|
| 1080 |
+
{
|
| 1081 |
+
"epoch": 0.36,
|
| 1082 |
+
"learning_rate": 9.646069213842769e-05,
|
| 1083 |
+
"loss": 0.7765,
|
| 1084 |
+
"step": 17800
|
| 1085 |
+
},
|
| 1086 |
+
{
|
| 1087 |
+
"epoch": 0.36,
|
| 1088 |
+
"learning_rate": 9.644068813762753e-05,
|
| 1089 |
+
"loss": 0.8268,
|
| 1090 |
+
"step": 17900
|
| 1091 |
+
},
|
| 1092 |
+
{
|
| 1093 |
+
"epoch": 0.36,
|
| 1094 |
+
"learning_rate": 9.642068413682736e-05,
|
| 1095 |
+
"loss": 0.7864,
|
| 1096 |
+
"step": 18000
|
| 1097 |
+
},
|
| 1098 |
+
{
|
| 1099 |
+
"epoch": 0.36,
|
| 1100 |
+
"learning_rate": 9.640068013602721e-05,
|
| 1101 |
+
"loss": 0.7741,
|
| 1102 |
+
"step": 18100
|
| 1103 |
+
},
|
| 1104 |
+
{
|
| 1105 |
+
"epoch": 0.36,
|
| 1106 |
+
"learning_rate": 9.638067613522705e-05,
|
| 1107 |
+
"loss": 0.831,
|
| 1108 |
+
"step": 18200
|
| 1109 |
+
},
|
| 1110 |
+
{
|
| 1111 |
+
"epoch": 0.37,
|
| 1112 |
+
"learning_rate": 9.636067213442688e-05,
|
| 1113 |
+
"loss": 0.7912,
|
| 1114 |
+
"step": 18300
|
| 1115 |
+
},
|
| 1116 |
+
{
|
| 1117 |
+
"epoch": 0.37,
|
| 1118 |
+
"learning_rate": 9.634066813362673e-05,
|
| 1119 |
+
"loss": 0.7864,
|
| 1120 |
+
"step": 18400
|
| 1121 |
+
},
|
| 1122 |
+
{
|
| 1123 |
+
"epoch": 0.37,
|
| 1124 |
+
"learning_rate": 9.632066413282656e-05,
|
| 1125 |
+
"loss": 0.7639,
|
| 1126 |
+
"step": 18500
|
| 1127 |
+
},
|
| 1128 |
+
{
|
| 1129 |
+
"epoch": 0.37,
|
| 1130 |
+
"learning_rate": 9.63006601320264e-05,
|
| 1131 |
+
"loss": 0.7874,
|
| 1132 |
+
"step": 18600
|
| 1133 |
+
},
|
| 1134 |
+
{
|
| 1135 |
+
"epoch": 0.37,
|
| 1136 |
+
"learning_rate": 9.628065613122625e-05,
|
| 1137 |
+
"loss": 0.7642,
|
| 1138 |
+
"step": 18700
|
| 1139 |
+
},
|
| 1140 |
+
{
|
| 1141 |
+
"epoch": 0.38,
|
| 1142 |
+
"learning_rate": 9.626065213042609e-05,
|
| 1143 |
+
"loss": 0.756,
|
| 1144 |
+
"step": 18800
|
| 1145 |
+
},
|
| 1146 |
+
{
|
| 1147 |
+
"epoch": 0.38,
|
| 1148 |
+
"learning_rate": 9.624064812962594e-05,
|
| 1149 |
+
"loss": 0.8036,
|
| 1150 |
+
"step": 18900
|
| 1151 |
+
},
|
| 1152 |
+
{
|
| 1153 |
+
"epoch": 0.38,
|
| 1154 |
+
"learning_rate": 9.622064412882577e-05,
|
| 1155 |
+
"loss": 0.7537,
|
| 1156 |
+
"step": 19000
|
| 1157 |
+
},
|
| 1158 |
+
{
|
| 1159 |
+
"epoch": 0.38,
|
| 1160 |
+
"learning_rate": 9.620064012802561e-05,
|
| 1161 |
+
"loss": 0.8193,
|
| 1162 |
+
"step": 19100
|
| 1163 |
+
},
|
| 1164 |
+
{
|
| 1165 |
+
"epoch": 0.38,
|
| 1166 |
+
"learning_rate": 9.618063612722546e-05,
|
| 1167 |
+
"loss": 0.8036,
|
| 1168 |
+
"step": 19200
|
| 1169 |
+
},
|
| 1170 |
+
{
|
| 1171 |
+
"epoch": 0.39,
|
| 1172 |
+
"learning_rate": 9.616063212642529e-05,
|
| 1173 |
+
"loss": 0.784,
|
| 1174 |
+
"step": 19300
|
| 1175 |
+
},
|
| 1176 |
+
{
|
| 1177 |
+
"epoch": 0.39,
|
| 1178 |
+
"learning_rate": 9.614062812562513e-05,
|
| 1179 |
+
"loss": 0.7861,
|
| 1180 |
+
"step": 19400
|
| 1181 |
+
},
|
| 1182 |
+
{
|
| 1183 |
+
"epoch": 0.39,
|
| 1184 |
+
"learning_rate": 9.612062412482496e-05,
|
| 1185 |
+
"loss": 0.7882,
|
| 1186 |
+
"step": 19500
|
| 1187 |
+
},
|
| 1188 |
+
{
|
| 1189 |
+
"epoch": 0.39,
|
| 1190 |
+
"learning_rate": 9.610062012402481e-05,
|
| 1191 |
+
"loss": 0.8048,
|
| 1192 |
+
"step": 19600
|
| 1193 |
+
},
|
| 1194 |
+
{
|
| 1195 |
+
"epoch": 0.39,
|
| 1196 |
+
"learning_rate": 9.608061612322465e-05,
|
| 1197 |
+
"loss": 0.8544,
|
| 1198 |
+
"step": 19700
|
| 1199 |
+
},
|
| 1200 |
+
{
|
| 1201 |
+
"epoch": 0.4,
|
| 1202 |
+
"learning_rate": 9.606061212242448e-05,
|
| 1203 |
+
"loss": 0.7827,
|
| 1204 |
+
"step": 19800
|
| 1205 |
+
},
|
| 1206 |
+
{
|
| 1207 |
+
"epoch": 0.4,
|
| 1208 |
+
"learning_rate": 9.604060812162433e-05,
|
| 1209 |
+
"loss": 0.7641,
|
| 1210 |
+
"step": 19900
|
| 1211 |
+
},
|
| 1212 |
+
{
|
| 1213 |
+
"epoch": 0.4,
|
| 1214 |
+
"learning_rate": 9.602060412082416e-05,
|
| 1215 |
+
"loss": 0.7146,
|
| 1216 |
+
"step": 20000
|
| 1217 |
+
},
|
| 1218 |
+
{
|
| 1219 |
+
"epoch": 0.4,
|
| 1220 |
+
"eval_loss": 0.7286937832832336,
|
| 1221 |
+
"eval_runtime": 689.1627,
|
| 1222 |
+
"eval_samples_per_second": 7.217,
|
| 1223 |
+
"eval_steps_per_second": 0.903,
|
| 1224 |
+
"step": 20000
|
| 1225 |
+
}
|
| 1226 |
+
],
|
| 1227 |
+
"max_steps": 500000,
|
| 1228 |
+
"num_train_epochs": 10,
|
| 1229 |
+
"total_flos": 2.798206695804764e+17,
|
| 1230 |
+
"trial_name": null,
|
| 1231 |
+
"trial_params": null
|
| 1232 |
+
}
|
checkpoint-20000/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:86e3e950534e1e3e08a50aef6d282bea64ab70d530ec27d5c1bcac285ba164ae
|
| 3 |
+
size 4408
|
checkpoint-30000/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<pad>": 32000
|
| 3 |
+
}
|
checkpoint-30000/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1b8a61e84b96b346aed5a0db359e918c51171457a1f7ba82cd8a215dd41ba8ce
|
| 3 |
+
size 134433530
|
checkpoint-30000/pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0ab38875812e802986bddfd3dbec386f4c37cbef696dbb0a125145bb82f7d30d
|
| 3 |
+
size 67201802
|
checkpoint-30000/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c45b7cfba8b08d90b8d62210e3fcfb44c2c6e6404a37cf43a5fe02972945d86c
|
| 3 |
+
size 14244
|
checkpoint-30000/scaler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e8f6138142907ab834f63d6fdce6932337833f54ae4f85014504cef61c25cc06
|
| 3 |
+
size 988
|
checkpoint-30000/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:fbf0ebc3e4d52a28a3ebf668b705ef0d647fefb862bd7a3407539a224342856c
|
| 3 |
+
size 1064
|
checkpoint-30000/special_tokens_map.json
ADDED
|
@@ -0,0 +1,24 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"content": "<s>",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": false,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"eos_token": {
|
| 10 |
+
"content": "</s>",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"pad_token": "<s>",
|
| 17 |
+
"unk_token": {
|
| 18 |
+
"content": "<unk>",
|
| 19 |
+
"lstrip": false,
|
| 20 |
+
"normalized": false,
|
| 21 |
+
"rstrip": false,
|
| 22 |
+
"single_word": false
|
| 23 |
+
}
|
| 24 |
+
}
|
checkpoint-30000/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
checkpoint-30000/tokenizer.model
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9e556afd44213b6bd1be2b850ebbbd98f5481437a8021afaf58ee7fb1818d347
|
| 3 |
+
size 499723
|
checkpoint-30000/tokenizer_config.json
ADDED
|
@@ -0,0 +1,34 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_bos_token": true,
|
| 3 |
+
"add_eos_token": false,
|
| 4 |
+
"bos_token": {
|
| 5 |
+
"__type": "AddedToken",
|
| 6 |
+
"content": "<s>",
|
| 7 |
+
"lstrip": false,
|
| 8 |
+
"normalized": false,
|
| 9 |
+
"rstrip": false,
|
| 10 |
+
"single_word": false
|
| 11 |
+
},
|
| 12 |
+
"clean_up_tokenization_spaces": false,
|
| 13 |
+
"eos_token": {
|
| 14 |
+
"__type": "AddedToken",
|
| 15 |
+
"content": "</s>",
|
| 16 |
+
"lstrip": false,
|
| 17 |
+
"normalized": false,
|
| 18 |
+
"rstrip": false,
|
| 19 |
+
"single_word": false
|
| 20 |
+
},
|
| 21 |
+
"legacy": false,
|
| 22 |
+
"model_max_length": 1000000000000000019884624838656,
|
| 23 |
+
"pad_token": null,
|
| 24 |
+
"sp_model_kwargs": {},
|
| 25 |
+
"tokenizer_class": "LlamaTokenizer",
|
| 26 |
+
"unk_token": {
|
| 27 |
+
"__type": "AddedToken",
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false
|
| 33 |
+
}
|
| 34 |
+
}
|
checkpoint-30000/trainer_state.json
ADDED
|
@@ -0,0 +1,1840 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.7202807068824768,
|
| 3 |
+
"best_model_checkpoint": "/media/xzuyn/c41d93ed-83a7-4577-9d38-5062b7e14438/xzuyn/Documents/AI_ROCM_5-6/LLaMa-LoRA-Tuner/data/lora_models/LLaMa-2-WizardLM-Uncensored-Tulu-Format-50K-7B-LoRA/checkpoint-30000",
|
| 4 |
+
"epoch": 0.6,
|
| 5 |
+
"global_step": 30000,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.0,
|
| 12 |
+
"learning_rate": 0.0001,
|
| 13 |
+
"loss": 1.1136,
|
| 14 |
+
"step": 100
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.0,
|
| 18 |
+
"learning_rate": 9.997999599919985e-05,
|
| 19 |
+
"loss": 0.9595,
|
| 20 |
+
"step": 200
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.01,
|
| 24 |
+
"learning_rate": 9.995999199839968e-05,
|
| 25 |
+
"loss": 0.8936,
|
| 26 |
+
"step": 300
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.01,
|
| 30 |
+
"learning_rate": 9.993998799759952e-05,
|
| 31 |
+
"loss": 0.9237,
|
| 32 |
+
"step": 400
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.01,
|
| 36 |
+
"learning_rate": 9.992038407681536e-05,
|
| 37 |
+
"loss": 0.9005,
|
| 38 |
+
"step": 500
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.01,
|
| 42 |
+
"learning_rate": 9.99003800760152e-05,
|
| 43 |
+
"loss": 0.8433,
|
| 44 |
+
"step": 600
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.01,
|
| 48 |
+
"learning_rate": 9.988037607521505e-05,
|
| 49 |
+
"loss": 0.863,
|
| 50 |
+
"step": 700
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.02,
|
| 54 |
+
"learning_rate": 9.98603720744149e-05,
|
| 55 |
+
"loss": 0.8634,
|
| 56 |
+
"step": 800
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.02,
|
| 60 |
+
"learning_rate": 9.984036807361474e-05,
|
| 61 |
+
"loss": 0.8463,
|
| 62 |
+
"step": 900
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.02,
|
| 66 |
+
"learning_rate": 9.982036407281457e-05,
|
| 67 |
+
"loss": 0.8527,
|
| 68 |
+
"step": 1000
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.02,
|
| 72 |
+
"learning_rate": 9.980036007201441e-05,
|
| 73 |
+
"loss": 0.8259,
|
| 74 |
+
"step": 1100
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.02,
|
| 78 |
+
"learning_rate": 9.978035607121424e-05,
|
| 79 |
+
"loss": 0.8473,
|
| 80 |
+
"step": 1200
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.03,
|
| 84 |
+
"learning_rate": 9.976035207041409e-05,
|
| 85 |
+
"loss": 0.8462,
|
| 86 |
+
"step": 1300
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.03,
|
| 90 |
+
"learning_rate": 9.974034806961393e-05,
|
| 91 |
+
"loss": 0.8609,
|
| 92 |
+
"step": 1400
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.03,
|
| 96 |
+
"learning_rate": 9.972034406881376e-05,
|
| 97 |
+
"loss": 0.8238,
|
| 98 |
+
"step": 1500
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.03,
|
| 102 |
+
"learning_rate": 9.970034006801361e-05,
|
| 103 |
+
"loss": 0.8212,
|
| 104 |
+
"step": 1600
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 0.03,
|
| 108 |
+
"learning_rate": 9.968033606721344e-05,
|
| 109 |
+
"loss": 0.8391,
|
| 110 |
+
"step": 1700
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.04,
|
| 114 |
+
"learning_rate": 9.966033206641328e-05,
|
| 115 |
+
"loss": 0.774,
|
| 116 |
+
"step": 1800
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 0.04,
|
| 120 |
+
"learning_rate": 9.964032806561313e-05,
|
| 121 |
+
"loss": 0.8149,
|
| 122 |
+
"step": 1900
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.04,
|
| 126 |
+
"learning_rate": 9.962032406481296e-05,
|
| 127 |
+
"loss": 0.8794,
|
| 128 |
+
"step": 2000
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.04,
|
| 132 |
+
"learning_rate": 9.96003200640128e-05,
|
| 133 |
+
"loss": 0.7747,
|
| 134 |
+
"step": 2100
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.04,
|
| 138 |
+
"learning_rate": 9.958031606321264e-05,
|
| 139 |
+
"loss": 0.8191,
|
| 140 |
+
"step": 2200
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 0.05,
|
| 144 |
+
"learning_rate": 9.956031206241248e-05,
|
| 145 |
+
"loss": 0.8219,
|
| 146 |
+
"step": 2300
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 0.05,
|
| 150 |
+
"learning_rate": 9.954030806161232e-05,
|
| 151 |
+
"loss": 0.8191,
|
| 152 |
+
"step": 2400
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 0.05,
|
| 156 |
+
"learning_rate": 9.952030406081217e-05,
|
| 157 |
+
"loss": 0.8662,
|
| 158 |
+
"step": 2500
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 0.05,
|
| 162 |
+
"learning_rate": 9.950030006001201e-05,
|
| 163 |
+
"loss": 0.787,
|
| 164 |
+
"step": 2600
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.05,
|
| 168 |
+
"learning_rate": 9.948029605921184e-05,
|
| 169 |
+
"loss": 0.8668,
|
| 170 |
+
"step": 2700
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.06,
|
| 174 |
+
"learning_rate": 9.946029205841169e-05,
|
| 175 |
+
"loss": 0.7817,
|
| 176 |
+
"step": 2800
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 0.06,
|
| 180 |
+
"learning_rate": 9.944028805761153e-05,
|
| 181 |
+
"loss": 0.8021,
|
| 182 |
+
"step": 2900
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 0.06,
|
| 186 |
+
"learning_rate": 9.942028405681136e-05,
|
| 187 |
+
"loss": 0.8324,
|
| 188 |
+
"step": 3000
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 0.06,
|
| 192 |
+
"learning_rate": 9.940028005601121e-05,
|
| 193 |
+
"loss": 0.8631,
|
| 194 |
+
"step": 3100
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 0.06,
|
| 198 |
+
"learning_rate": 9.938027605521105e-05,
|
| 199 |
+
"loss": 0.797,
|
| 200 |
+
"step": 3200
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 0.07,
|
| 204 |
+
"learning_rate": 9.936027205441088e-05,
|
| 205 |
+
"loss": 0.8224,
|
| 206 |
+
"step": 3300
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.07,
|
| 210 |
+
"learning_rate": 9.934026805361073e-05,
|
| 211 |
+
"loss": 0.8164,
|
| 212 |
+
"step": 3400
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.07,
|
| 216 |
+
"learning_rate": 9.932026405281056e-05,
|
| 217 |
+
"loss": 0.7975,
|
| 218 |
+
"step": 3500
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 0.07,
|
| 222 |
+
"learning_rate": 9.930046009201841e-05,
|
| 223 |
+
"loss": 0.8468,
|
| 224 |
+
"step": 3600
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 0.07,
|
| 228 |
+
"learning_rate": 9.928045609121824e-05,
|
| 229 |
+
"loss": 0.8169,
|
| 230 |
+
"step": 3700
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 0.08,
|
| 234 |
+
"learning_rate": 9.926045209041809e-05,
|
| 235 |
+
"loss": 0.7761,
|
| 236 |
+
"step": 3800
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 0.08,
|
| 240 |
+
"learning_rate": 9.924044808961793e-05,
|
| 241 |
+
"loss": 0.8042,
|
| 242 |
+
"step": 3900
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 0.08,
|
| 246 |
+
"learning_rate": 9.922044408881776e-05,
|
| 247 |
+
"loss": 0.8146,
|
| 248 |
+
"step": 4000
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 0.08,
|
| 252 |
+
"learning_rate": 9.920044008801761e-05,
|
| 253 |
+
"loss": 0.8243,
|
| 254 |
+
"step": 4100
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.08,
|
| 258 |
+
"learning_rate": 9.918043608721744e-05,
|
| 259 |
+
"loss": 0.8315,
|
| 260 |
+
"step": 4200
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 0.09,
|
| 264 |
+
"learning_rate": 9.916043208641728e-05,
|
| 265 |
+
"loss": 0.8014,
|
| 266 |
+
"step": 4300
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 0.09,
|
| 270 |
+
"learning_rate": 9.914042808561713e-05,
|
| 271 |
+
"loss": 0.8241,
|
| 272 |
+
"step": 4400
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 0.09,
|
| 276 |
+
"learning_rate": 9.912042408481696e-05,
|
| 277 |
+
"loss": 0.8233,
|
| 278 |
+
"step": 4500
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 0.09,
|
| 282 |
+
"learning_rate": 9.91004200840168e-05,
|
| 283 |
+
"loss": 0.7748,
|
| 284 |
+
"step": 4600
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 0.09,
|
| 288 |
+
"learning_rate": 9.908041608321665e-05,
|
| 289 |
+
"loss": 0.8197,
|
| 290 |
+
"step": 4700
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 0.1,
|
| 294 |
+
"learning_rate": 9.906041208241649e-05,
|
| 295 |
+
"loss": 0.8195,
|
| 296 |
+
"step": 4800
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.1,
|
| 300 |
+
"learning_rate": 9.904040808161634e-05,
|
| 301 |
+
"loss": 0.7864,
|
| 302 |
+
"step": 4900
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 0.1,
|
| 306 |
+
"learning_rate": 9.902040408081617e-05,
|
| 307 |
+
"loss": 0.7812,
|
| 308 |
+
"step": 5000
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 0.1,
|
| 312 |
+
"learning_rate": 9.900040008001601e-05,
|
| 313 |
+
"loss": 0.7966,
|
| 314 |
+
"step": 5100
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 0.1,
|
| 318 |
+
"learning_rate": 9.898039607921584e-05,
|
| 319 |
+
"loss": 0.793,
|
| 320 |
+
"step": 5200
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 0.11,
|
| 324 |
+
"learning_rate": 9.896039207841569e-05,
|
| 325 |
+
"loss": 0.8568,
|
| 326 |
+
"step": 5300
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 0.11,
|
| 330 |
+
"learning_rate": 9.894038807761553e-05,
|
| 331 |
+
"loss": 0.7614,
|
| 332 |
+
"step": 5400
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 0.11,
|
| 336 |
+
"learning_rate": 9.892038407681536e-05,
|
| 337 |
+
"loss": 0.8171,
|
| 338 |
+
"step": 5500
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.11,
|
| 342 |
+
"learning_rate": 9.890038007601521e-05,
|
| 343 |
+
"loss": 0.7649,
|
| 344 |
+
"step": 5600
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 0.11,
|
| 348 |
+
"learning_rate": 9.888037607521505e-05,
|
| 349 |
+
"loss": 0.8074,
|
| 350 |
+
"step": 5700
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 0.12,
|
| 354 |
+
"learning_rate": 9.886037207441488e-05,
|
| 355 |
+
"loss": 0.7591,
|
| 356 |
+
"step": 5800
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 0.12,
|
| 360 |
+
"learning_rate": 9.884036807361473e-05,
|
| 361 |
+
"loss": 0.8102,
|
| 362 |
+
"step": 5900
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 0.12,
|
| 366 |
+
"learning_rate": 9.882036407281456e-05,
|
| 367 |
+
"loss": 0.784,
|
| 368 |
+
"step": 6000
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 0.12,
|
| 372 |
+
"learning_rate": 9.88003600720144e-05,
|
| 373 |
+
"loss": 0.8367,
|
| 374 |
+
"step": 6100
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 0.12,
|
| 378 |
+
"learning_rate": 9.878035607121425e-05,
|
| 379 |
+
"loss": 0.8418,
|
| 380 |
+
"step": 6200
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.13,
|
| 384 |
+
"learning_rate": 9.876035207041409e-05,
|
| 385 |
+
"loss": 0.798,
|
| 386 |
+
"step": 6300
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 0.13,
|
| 390 |
+
"learning_rate": 9.874034806961394e-05,
|
| 391 |
+
"loss": 0.7928,
|
| 392 |
+
"step": 6400
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 0.13,
|
| 396 |
+
"learning_rate": 9.872034406881377e-05,
|
| 397 |
+
"loss": 0.8262,
|
| 398 |
+
"step": 6500
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 0.13,
|
| 402 |
+
"learning_rate": 9.870034006801361e-05,
|
| 403 |
+
"loss": 0.7965,
|
| 404 |
+
"step": 6600
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 0.13,
|
| 408 |
+
"learning_rate": 9.868033606721346e-05,
|
| 409 |
+
"loss": 0.7952,
|
| 410 |
+
"step": 6700
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 0.14,
|
| 414 |
+
"learning_rate": 9.866033206641329e-05,
|
| 415 |
+
"loss": 0.7987,
|
| 416 |
+
"step": 6800
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 0.14,
|
| 420 |
+
"learning_rate": 9.864032806561313e-05,
|
| 421 |
+
"loss": 0.7939,
|
| 422 |
+
"step": 6900
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.14,
|
| 426 |
+
"learning_rate": 9.862032406481296e-05,
|
| 427 |
+
"loss": 0.7484,
|
| 428 |
+
"step": 7000
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 0.14,
|
| 432 |
+
"learning_rate": 9.860032006401281e-05,
|
| 433 |
+
"loss": 0.819,
|
| 434 |
+
"step": 7100
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 0.14,
|
| 438 |
+
"learning_rate": 9.858031606321265e-05,
|
| 439 |
+
"loss": 0.747,
|
| 440 |
+
"step": 7200
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 0.15,
|
| 444 |
+
"learning_rate": 9.856031206241248e-05,
|
| 445 |
+
"loss": 0.8029,
|
| 446 |
+
"step": 7300
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 0.15,
|
| 450 |
+
"learning_rate": 9.854030806161233e-05,
|
| 451 |
+
"loss": 0.9013,
|
| 452 |
+
"step": 7400
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 0.15,
|
| 456 |
+
"learning_rate": 9.852030406081216e-05,
|
| 457 |
+
"loss": 0.8337,
|
| 458 |
+
"step": 7500
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 0.15,
|
| 462 |
+
"learning_rate": 9.8500300060012e-05,
|
| 463 |
+
"loss": 0.7927,
|
| 464 |
+
"step": 7600
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.15,
|
| 468 |
+
"learning_rate": 9.848029605921185e-05,
|
| 469 |
+
"loss": 0.8725,
|
| 470 |
+
"step": 7700
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 0.16,
|
| 474 |
+
"learning_rate": 9.846029205841169e-05,
|
| 475 |
+
"loss": 0.7865,
|
| 476 |
+
"step": 7800
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 0.16,
|
| 480 |
+
"learning_rate": 9.844028805761154e-05,
|
| 481 |
+
"loss": 0.7726,
|
| 482 |
+
"step": 7900
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 0.16,
|
| 486 |
+
"learning_rate": 9.842048409681936e-05,
|
| 487 |
+
"loss": 0.7868,
|
| 488 |
+
"step": 8000
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 0.16,
|
| 492 |
+
"learning_rate": 9.840048009601921e-05,
|
| 493 |
+
"loss": 0.7918,
|
| 494 |
+
"step": 8100
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 0.16,
|
| 498 |
+
"learning_rate": 9.838047609521905e-05,
|
| 499 |
+
"loss": 0.7924,
|
| 500 |
+
"step": 8200
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 0.17,
|
| 504 |
+
"learning_rate": 9.836047209441888e-05,
|
| 505 |
+
"loss": 0.8036,
|
| 506 |
+
"step": 8300
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 0.17,
|
| 510 |
+
"learning_rate": 9.834046809361873e-05,
|
| 511 |
+
"loss": 0.7844,
|
| 512 |
+
"step": 8400
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 0.17,
|
| 516 |
+
"learning_rate": 9.832046409281857e-05,
|
| 517 |
+
"loss": 0.8124,
|
| 518 |
+
"step": 8500
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 0.17,
|
| 522 |
+
"learning_rate": 9.830046009201842e-05,
|
| 523 |
+
"loss": 0.837,
|
| 524 |
+
"step": 8600
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 0.17,
|
| 528 |
+
"learning_rate": 9.828045609121826e-05,
|
| 529 |
+
"loss": 0.8315,
|
| 530 |
+
"step": 8700
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 0.18,
|
| 534 |
+
"learning_rate": 9.826045209041809e-05,
|
| 535 |
+
"loss": 0.7738,
|
| 536 |
+
"step": 8800
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 0.18,
|
| 540 |
+
"learning_rate": 9.824044808961794e-05,
|
| 541 |
+
"loss": 0.8164,
|
| 542 |
+
"step": 8900
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 0.18,
|
| 546 |
+
"learning_rate": 9.822044408881777e-05,
|
| 547 |
+
"loss": 0.7529,
|
| 548 |
+
"step": 9000
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 0.18,
|
| 552 |
+
"learning_rate": 9.820044008801761e-05,
|
| 553 |
+
"loss": 0.8037,
|
| 554 |
+
"step": 9100
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 0.18,
|
| 558 |
+
"learning_rate": 9.818043608721746e-05,
|
| 559 |
+
"loss": 0.8537,
|
| 560 |
+
"step": 9200
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 0.19,
|
| 564 |
+
"learning_rate": 9.816043208641729e-05,
|
| 565 |
+
"loss": 0.8241,
|
| 566 |
+
"step": 9300
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 0.19,
|
| 570 |
+
"learning_rate": 9.814042808561713e-05,
|
| 571 |
+
"loss": 0.7729,
|
| 572 |
+
"step": 9400
|
| 573 |
+
},
|
| 574 |
+
{
|
| 575 |
+
"epoch": 0.19,
|
| 576 |
+
"learning_rate": 9.812042408481696e-05,
|
| 577 |
+
"loss": 0.8116,
|
| 578 |
+
"step": 9500
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 0.19,
|
| 582 |
+
"learning_rate": 9.81004200840168e-05,
|
| 583 |
+
"loss": 0.8254,
|
| 584 |
+
"step": 9600
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 0.19,
|
| 588 |
+
"learning_rate": 9.808041608321665e-05,
|
| 589 |
+
"loss": 0.8285,
|
| 590 |
+
"step": 9700
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 0.2,
|
| 594 |
+
"learning_rate": 9.806041208241648e-05,
|
| 595 |
+
"loss": 0.7577,
|
| 596 |
+
"step": 9800
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 0.2,
|
| 600 |
+
"learning_rate": 9.804040808161633e-05,
|
| 601 |
+
"loss": 0.8024,
|
| 602 |
+
"step": 9900
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"epoch": 0.2,
|
| 606 |
+
"learning_rate": 9.802040408081616e-05,
|
| 607 |
+
"loss": 0.8078,
|
| 608 |
+
"step": 10000
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 0.2,
|
| 612 |
+
"eval_loss": 0.740916907787323,
|
| 613 |
+
"eval_runtime": 689.7578,
|
| 614 |
+
"eval_samples_per_second": 7.211,
|
| 615 |
+
"eval_steps_per_second": 0.902,
|
| 616 |
+
"step": 10000
|
| 617 |
+
},
|
| 618 |
+
{
|
| 619 |
+
"epoch": 0.2,
|
| 620 |
+
"learning_rate": 9.8000400080016e-05,
|
| 621 |
+
"loss": 0.7849,
|
| 622 |
+
"step": 10100
|
| 623 |
+
},
|
| 624 |
+
{
|
| 625 |
+
"epoch": 0.2,
|
| 626 |
+
"learning_rate": 9.798039607921585e-05,
|
| 627 |
+
"loss": 0.7931,
|
| 628 |
+
"step": 10200
|
| 629 |
+
},
|
| 630 |
+
{
|
| 631 |
+
"epoch": 0.21,
|
| 632 |
+
"learning_rate": 9.796039207841569e-05,
|
| 633 |
+
"loss": 0.7884,
|
| 634 |
+
"step": 10300
|
| 635 |
+
},
|
| 636 |
+
{
|
| 637 |
+
"epoch": 0.21,
|
| 638 |
+
"learning_rate": 9.794038807761554e-05,
|
| 639 |
+
"loss": 0.7949,
|
| 640 |
+
"step": 10400
|
| 641 |
+
},
|
| 642 |
+
{
|
| 643 |
+
"epoch": 0.21,
|
| 644 |
+
"learning_rate": 9.792038407681537e-05,
|
| 645 |
+
"loss": 0.7916,
|
| 646 |
+
"step": 10500
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"epoch": 0.21,
|
| 650 |
+
"learning_rate": 9.790038007601521e-05,
|
| 651 |
+
"loss": 0.8072,
|
| 652 |
+
"step": 10600
|
| 653 |
+
},
|
| 654 |
+
{
|
| 655 |
+
"epoch": 0.21,
|
| 656 |
+
"learning_rate": 9.788037607521506e-05,
|
| 657 |
+
"loss": 0.7965,
|
| 658 |
+
"step": 10700
|
| 659 |
+
},
|
| 660 |
+
{
|
| 661 |
+
"epoch": 0.22,
|
| 662 |
+
"learning_rate": 9.786037207441489e-05,
|
| 663 |
+
"loss": 0.8039,
|
| 664 |
+
"step": 10800
|
| 665 |
+
},
|
| 666 |
+
{
|
| 667 |
+
"epoch": 0.22,
|
| 668 |
+
"learning_rate": 9.784036807361473e-05,
|
| 669 |
+
"loss": 0.8011,
|
| 670 |
+
"step": 10900
|
| 671 |
+
},
|
| 672 |
+
{
|
| 673 |
+
"epoch": 0.22,
|
| 674 |
+
"learning_rate": 9.782036407281456e-05,
|
| 675 |
+
"loss": 0.8014,
|
| 676 |
+
"step": 11000
|
| 677 |
+
},
|
| 678 |
+
{
|
| 679 |
+
"epoch": 0.22,
|
| 680 |
+
"learning_rate": 9.78003600720144e-05,
|
| 681 |
+
"loss": 0.7835,
|
| 682 |
+
"step": 11100
|
| 683 |
+
},
|
| 684 |
+
{
|
| 685 |
+
"epoch": 0.22,
|
| 686 |
+
"learning_rate": 9.778035607121425e-05,
|
| 687 |
+
"loss": 0.8469,
|
| 688 |
+
"step": 11200
|
| 689 |
+
},
|
| 690 |
+
{
|
| 691 |
+
"epoch": 0.23,
|
| 692 |
+
"learning_rate": 9.776035207041408e-05,
|
| 693 |
+
"loss": 0.8018,
|
| 694 |
+
"step": 11300
|
| 695 |
+
},
|
| 696 |
+
{
|
| 697 |
+
"epoch": 0.23,
|
| 698 |
+
"learning_rate": 9.774034806961393e-05,
|
| 699 |
+
"loss": 0.7839,
|
| 700 |
+
"step": 11400
|
| 701 |
+
},
|
| 702 |
+
{
|
| 703 |
+
"epoch": 0.23,
|
| 704 |
+
"learning_rate": 9.772034406881376e-05,
|
| 705 |
+
"loss": 0.8312,
|
| 706 |
+
"step": 11500
|
| 707 |
+
},
|
| 708 |
+
{
|
| 709 |
+
"epoch": 0.23,
|
| 710 |
+
"learning_rate": 9.770054010802161e-05,
|
| 711 |
+
"loss": 0.7886,
|
| 712 |
+
"step": 11600
|
| 713 |
+
},
|
| 714 |
+
{
|
| 715 |
+
"epoch": 0.23,
|
| 716 |
+
"learning_rate": 9.768053610722146e-05,
|
| 717 |
+
"loss": 0.8096,
|
| 718 |
+
"step": 11700
|
| 719 |
+
},
|
| 720 |
+
{
|
| 721 |
+
"epoch": 0.24,
|
| 722 |
+
"learning_rate": 9.766053210642129e-05,
|
| 723 |
+
"loss": 0.7806,
|
| 724 |
+
"step": 11800
|
| 725 |
+
},
|
| 726 |
+
{
|
| 727 |
+
"epoch": 0.24,
|
| 728 |
+
"learning_rate": 9.764052810562113e-05,
|
| 729 |
+
"loss": 0.8512,
|
| 730 |
+
"step": 11900
|
| 731 |
+
},
|
| 732 |
+
{
|
| 733 |
+
"epoch": 0.24,
|
| 734 |
+
"learning_rate": 9.762052410482096e-05,
|
| 735 |
+
"loss": 0.8475,
|
| 736 |
+
"step": 12000
|
| 737 |
+
},
|
| 738 |
+
{
|
| 739 |
+
"epoch": 0.24,
|
| 740 |
+
"learning_rate": 9.76005201040208e-05,
|
| 741 |
+
"loss": 0.749,
|
| 742 |
+
"step": 12100
|
| 743 |
+
},
|
| 744 |
+
{
|
| 745 |
+
"epoch": 0.24,
|
| 746 |
+
"learning_rate": 9.758051610322065e-05,
|
| 747 |
+
"loss": 0.8245,
|
| 748 |
+
"step": 12200
|
| 749 |
+
},
|
| 750 |
+
{
|
| 751 |
+
"epoch": 0.25,
|
| 752 |
+
"learning_rate": 9.756051210242048e-05,
|
| 753 |
+
"loss": 0.7966,
|
| 754 |
+
"step": 12300
|
| 755 |
+
},
|
| 756 |
+
{
|
| 757 |
+
"epoch": 0.25,
|
| 758 |
+
"learning_rate": 9.754050810162033e-05,
|
| 759 |
+
"loss": 0.773,
|
| 760 |
+
"step": 12400
|
| 761 |
+
},
|
| 762 |
+
{
|
| 763 |
+
"epoch": 0.25,
|
| 764 |
+
"learning_rate": 9.752050410082017e-05,
|
| 765 |
+
"loss": 0.7876,
|
| 766 |
+
"step": 12500
|
| 767 |
+
},
|
| 768 |
+
{
|
| 769 |
+
"epoch": 0.25,
|
| 770 |
+
"learning_rate": 9.750050010002001e-05,
|
| 771 |
+
"loss": 0.7982,
|
| 772 |
+
"step": 12600
|
| 773 |
+
},
|
| 774 |
+
{
|
| 775 |
+
"epoch": 0.25,
|
| 776 |
+
"learning_rate": 9.748049609921986e-05,
|
| 777 |
+
"loss": 0.7509,
|
| 778 |
+
"step": 12700
|
| 779 |
+
},
|
| 780 |
+
{
|
| 781 |
+
"epoch": 0.26,
|
| 782 |
+
"learning_rate": 9.746049209841969e-05,
|
| 783 |
+
"loss": 0.7815,
|
| 784 |
+
"step": 12800
|
| 785 |
+
},
|
| 786 |
+
{
|
| 787 |
+
"epoch": 0.26,
|
| 788 |
+
"learning_rate": 9.744048809761953e-05,
|
| 789 |
+
"loss": 0.8516,
|
| 790 |
+
"step": 12900
|
| 791 |
+
},
|
| 792 |
+
{
|
| 793 |
+
"epoch": 0.26,
|
| 794 |
+
"learning_rate": 9.742048409681937e-05,
|
| 795 |
+
"loss": 0.7684,
|
| 796 |
+
"step": 13000
|
| 797 |
+
},
|
| 798 |
+
{
|
| 799 |
+
"epoch": 0.26,
|
| 800 |
+
"learning_rate": 9.740048009601921e-05,
|
| 801 |
+
"loss": 0.7889,
|
| 802 |
+
"step": 13100
|
| 803 |
+
},
|
| 804 |
+
{
|
| 805 |
+
"epoch": 0.26,
|
| 806 |
+
"learning_rate": 9.738047609521905e-05,
|
| 807 |
+
"loss": 0.7882,
|
| 808 |
+
"step": 13200
|
| 809 |
+
},
|
| 810 |
+
{
|
| 811 |
+
"epoch": 0.27,
|
| 812 |
+
"learning_rate": 9.736047209441889e-05,
|
| 813 |
+
"loss": 0.8196,
|
| 814 |
+
"step": 13300
|
| 815 |
+
},
|
| 816 |
+
{
|
| 817 |
+
"epoch": 0.27,
|
| 818 |
+
"learning_rate": 9.734046809361873e-05,
|
| 819 |
+
"loss": 0.7889,
|
| 820 |
+
"step": 13400
|
| 821 |
+
},
|
| 822 |
+
{
|
| 823 |
+
"epoch": 0.27,
|
| 824 |
+
"learning_rate": 9.732046409281856e-05,
|
| 825 |
+
"loss": 0.7931,
|
| 826 |
+
"step": 13500
|
| 827 |
+
},
|
| 828 |
+
{
|
| 829 |
+
"epoch": 0.27,
|
| 830 |
+
"learning_rate": 9.73004600920184e-05,
|
| 831 |
+
"loss": 0.8172,
|
| 832 |
+
"step": 13600
|
| 833 |
+
},
|
| 834 |
+
{
|
| 835 |
+
"epoch": 0.27,
|
| 836 |
+
"learning_rate": 9.728045609121825e-05,
|
| 837 |
+
"loss": 0.8364,
|
| 838 |
+
"step": 13700
|
| 839 |
+
},
|
| 840 |
+
{
|
| 841 |
+
"epoch": 0.28,
|
| 842 |
+
"learning_rate": 9.726045209041808e-05,
|
| 843 |
+
"loss": 0.822,
|
| 844 |
+
"step": 13800
|
| 845 |
+
},
|
| 846 |
+
{
|
| 847 |
+
"epoch": 0.28,
|
| 848 |
+
"learning_rate": 9.724044808961793e-05,
|
| 849 |
+
"loss": 0.8546,
|
| 850 |
+
"step": 13900
|
| 851 |
+
},
|
| 852 |
+
{
|
| 853 |
+
"epoch": 0.28,
|
| 854 |
+
"learning_rate": 9.722044408881777e-05,
|
| 855 |
+
"loss": 0.7977,
|
| 856 |
+
"step": 14000
|
| 857 |
+
},
|
| 858 |
+
{
|
| 859 |
+
"epoch": 0.28,
|
| 860 |
+
"learning_rate": 9.720044008801761e-05,
|
| 861 |
+
"loss": 0.7591,
|
| 862 |
+
"step": 14100
|
| 863 |
+
},
|
| 864 |
+
{
|
| 865 |
+
"epoch": 0.28,
|
| 866 |
+
"learning_rate": 9.718063612722545e-05,
|
| 867 |
+
"loss": 0.7787,
|
| 868 |
+
"step": 14200
|
| 869 |
+
},
|
| 870 |
+
{
|
| 871 |
+
"epoch": 0.29,
|
| 872 |
+
"learning_rate": 9.716063212642529e-05,
|
| 873 |
+
"loss": 0.8361,
|
| 874 |
+
"step": 14300
|
| 875 |
+
},
|
| 876 |
+
{
|
| 877 |
+
"epoch": 0.29,
|
| 878 |
+
"learning_rate": 9.714062812562513e-05,
|
| 879 |
+
"loss": 0.8153,
|
| 880 |
+
"step": 14400
|
| 881 |
+
},
|
| 882 |
+
{
|
| 883 |
+
"epoch": 0.29,
|
| 884 |
+
"learning_rate": 9.712062412482496e-05,
|
| 885 |
+
"loss": 0.776,
|
| 886 |
+
"step": 14500
|
| 887 |
+
},
|
| 888 |
+
{
|
| 889 |
+
"epoch": 0.29,
|
| 890 |
+
"learning_rate": 9.71006201240248e-05,
|
| 891 |
+
"loss": 0.7491,
|
| 892 |
+
"step": 14600
|
| 893 |
+
},
|
| 894 |
+
{
|
| 895 |
+
"epoch": 0.29,
|
| 896 |
+
"learning_rate": 9.708061612322465e-05,
|
| 897 |
+
"loss": 0.7975,
|
| 898 |
+
"step": 14700
|
| 899 |
+
},
|
| 900 |
+
{
|
| 901 |
+
"epoch": 0.3,
|
| 902 |
+
"learning_rate": 9.70606121224245e-05,
|
| 903 |
+
"loss": 0.7637,
|
| 904 |
+
"step": 14800
|
| 905 |
+
},
|
| 906 |
+
{
|
| 907 |
+
"epoch": 0.3,
|
| 908 |
+
"learning_rate": 9.704060812162434e-05,
|
| 909 |
+
"loss": 0.7977,
|
| 910 |
+
"step": 14900
|
| 911 |
+
},
|
| 912 |
+
{
|
| 913 |
+
"epoch": 0.3,
|
| 914 |
+
"learning_rate": 9.702060412082417e-05,
|
| 915 |
+
"loss": 0.8041,
|
| 916 |
+
"step": 15000
|
| 917 |
+
},
|
| 918 |
+
{
|
| 919 |
+
"epoch": 0.3,
|
| 920 |
+
"learning_rate": 9.700060012002401e-05,
|
| 921 |
+
"loss": 0.7997,
|
| 922 |
+
"step": 15100
|
| 923 |
+
},
|
| 924 |
+
{
|
| 925 |
+
"epoch": 0.3,
|
| 926 |
+
"learning_rate": 9.698059611922386e-05,
|
| 927 |
+
"loss": 0.8221,
|
| 928 |
+
"step": 15200
|
| 929 |
+
},
|
| 930 |
+
{
|
| 931 |
+
"epoch": 0.31,
|
| 932 |
+
"learning_rate": 9.696059211842369e-05,
|
| 933 |
+
"loss": 0.7657,
|
| 934 |
+
"step": 15300
|
| 935 |
+
},
|
| 936 |
+
{
|
| 937 |
+
"epoch": 0.31,
|
| 938 |
+
"learning_rate": 9.694058811762353e-05,
|
| 939 |
+
"loss": 0.8091,
|
| 940 |
+
"step": 15400
|
| 941 |
+
},
|
| 942 |
+
{
|
| 943 |
+
"epoch": 0.31,
|
| 944 |
+
"learning_rate": 9.692058411682336e-05,
|
| 945 |
+
"loss": 0.7692,
|
| 946 |
+
"step": 15500
|
| 947 |
+
},
|
| 948 |
+
{
|
| 949 |
+
"epoch": 0.31,
|
| 950 |
+
"learning_rate": 9.690058011602321e-05,
|
| 951 |
+
"loss": 0.7988,
|
| 952 |
+
"step": 15600
|
| 953 |
+
},
|
| 954 |
+
{
|
| 955 |
+
"epoch": 0.31,
|
| 956 |
+
"learning_rate": 9.688057611522305e-05,
|
| 957 |
+
"loss": 0.7457,
|
| 958 |
+
"step": 15700
|
| 959 |
+
},
|
| 960 |
+
{
|
| 961 |
+
"epoch": 0.32,
|
| 962 |
+
"learning_rate": 9.686057211442288e-05,
|
| 963 |
+
"loss": 0.8223,
|
| 964 |
+
"step": 15800
|
| 965 |
+
},
|
| 966 |
+
{
|
| 967 |
+
"epoch": 0.32,
|
| 968 |
+
"learning_rate": 9.684056811362273e-05,
|
| 969 |
+
"loss": 0.8366,
|
| 970 |
+
"step": 15900
|
| 971 |
+
},
|
| 972 |
+
{
|
| 973 |
+
"epoch": 0.32,
|
| 974 |
+
"learning_rate": 9.682056411282256e-05,
|
| 975 |
+
"loss": 0.7959,
|
| 976 |
+
"step": 16000
|
| 977 |
+
},
|
| 978 |
+
{
|
| 979 |
+
"epoch": 0.32,
|
| 980 |
+
"learning_rate": 9.680076015203041e-05,
|
| 981 |
+
"loss": 0.7944,
|
| 982 |
+
"step": 16100
|
| 983 |
+
},
|
| 984 |
+
{
|
| 985 |
+
"epoch": 0.32,
|
| 986 |
+
"learning_rate": 9.678075615123024e-05,
|
| 987 |
+
"loss": 0.7645,
|
| 988 |
+
"step": 16200
|
| 989 |
+
},
|
| 990 |
+
{
|
| 991 |
+
"epoch": 0.33,
|
| 992 |
+
"learning_rate": 9.676075215043009e-05,
|
| 993 |
+
"loss": 0.8111,
|
| 994 |
+
"step": 16300
|
| 995 |
+
},
|
| 996 |
+
{
|
| 997 |
+
"epoch": 0.33,
|
| 998 |
+
"learning_rate": 9.674074814962993e-05,
|
| 999 |
+
"loss": 0.7907,
|
| 1000 |
+
"step": 16400
|
| 1001 |
+
},
|
| 1002 |
+
{
|
| 1003 |
+
"epoch": 0.33,
|
| 1004 |
+
"learning_rate": 9.672074414882976e-05,
|
| 1005 |
+
"loss": 0.8108,
|
| 1006 |
+
"step": 16500
|
| 1007 |
+
},
|
| 1008 |
+
{
|
| 1009 |
+
"epoch": 0.33,
|
| 1010 |
+
"learning_rate": 9.670074014802961e-05,
|
| 1011 |
+
"loss": 0.8119,
|
| 1012 |
+
"step": 16600
|
| 1013 |
+
},
|
| 1014 |
+
{
|
| 1015 |
+
"epoch": 0.33,
|
| 1016 |
+
"learning_rate": 9.668073614722945e-05,
|
| 1017 |
+
"loss": 0.8109,
|
| 1018 |
+
"step": 16700
|
| 1019 |
+
},
|
| 1020 |
+
{
|
| 1021 |
+
"epoch": 0.34,
|
| 1022 |
+
"learning_rate": 9.666073214642928e-05,
|
| 1023 |
+
"loss": 0.7862,
|
| 1024 |
+
"step": 16800
|
| 1025 |
+
},
|
| 1026 |
+
{
|
| 1027 |
+
"epoch": 0.34,
|
| 1028 |
+
"learning_rate": 9.664072814562913e-05,
|
| 1029 |
+
"loss": 0.7886,
|
| 1030 |
+
"step": 16900
|
| 1031 |
+
},
|
| 1032 |
+
{
|
| 1033 |
+
"epoch": 0.34,
|
| 1034 |
+
"learning_rate": 9.662072414482897e-05,
|
| 1035 |
+
"loss": 0.786,
|
| 1036 |
+
"step": 17000
|
| 1037 |
+
},
|
| 1038 |
+
{
|
| 1039 |
+
"epoch": 0.34,
|
| 1040 |
+
"learning_rate": 9.660072014402882e-05,
|
| 1041 |
+
"loss": 0.8014,
|
| 1042 |
+
"step": 17100
|
| 1043 |
+
},
|
| 1044 |
+
{
|
| 1045 |
+
"epoch": 0.34,
|
| 1046 |
+
"learning_rate": 9.658071614322865e-05,
|
| 1047 |
+
"loss": 0.7486,
|
| 1048 |
+
"step": 17200
|
| 1049 |
+
},
|
| 1050 |
+
{
|
| 1051 |
+
"epoch": 0.35,
|
| 1052 |
+
"learning_rate": 9.65607121424285e-05,
|
| 1053 |
+
"loss": 0.7688,
|
| 1054 |
+
"step": 17300
|
| 1055 |
+
},
|
| 1056 |
+
{
|
| 1057 |
+
"epoch": 0.35,
|
| 1058 |
+
"learning_rate": 9.654070814162834e-05,
|
| 1059 |
+
"loss": 0.7815,
|
| 1060 |
+
"step": 17400
|
| 1061 |
+
},
|
| 1062 |
+
{
|
| 1063 |
+
"epoch": 0.35,
|
| 1064 |
+
"learning_rate": 9.652070414082817e-05,
|
| 1065 |
+
"loss": 0.7857,
|
| 1066 |
+
"step": 17500
|
| 1067 |
+
},
|
| 1068 |
+
{
|
| 1069 |
+
"epoch": 0.35,
|
| 1070 |
+
"learning_rate": 9.650070014002801e-05,
|
| 1071 |
+
"loss": 0.777,
|
| 1072 |
+
"step": 17600
|
| 1073 |
+
},
|
| 1074 |
+
{
|
| 1075 |
+
"epoch": 0.35,
|
| 1076 |
+
"learning_rate": 9.648069613922786e-05,
|
| 1077 |
+
"loss": 0.8186,
|
| 1078 |
+
"step": 17700
|
| 1079 |
+
},
|
| 1080 |
+
{
|
| 1081 |
+
"epoch": 0.36,
|
| 1082 |
+
"learning_rate": 9.646069213842769e-05,
|
| 1083 |
+
"loss": 0.7765,
|
| 1084 |
+
"step": 17800
|
| 1085 |
+
},
|
| 1086 |
+
{
|
| 1087 |
+
"epoch": 0.36,
|
| 1088 |
+
"learning_rate": 9.644068813762753e-05,
|
| 1089 |
+
"loss": 0.8268,
|
| 1090 |
+
"step": 17900
|
| 1091 |
+
},
|
| 1092 |
+
{
|
| 1093 |
+
"epoch": 0.36,
|
| 1094 |
+
"learning_rate": 9.642068413682736e-05,
|
| 1095 |
+
"loss": 0.7864,
|
| 1096 |
+
"step": 18000
|
| 1097 |
+
},
|
| 1098 |
+
{
|
| 1099 |
+
"epoch": 0.36,
|
| 1100 |
+
"learning_rate": 9.640068013602721e-05,
|
| 1101 |
+
"loss": 0.7741,
|
| 1102 |
+
"step": 18100
|
| 1103 |
+
},
|
| 1104 |
+
{
|
| 1105 |
+
"epoch": 0.36,
|
| 1106 |
+
"learning_rate": 9.638067613522705e-05,
|
| 1107 |
+
"loss": 0.831,
|
| 1108 |
+
"step": 18200
|
| 1109 |
+
},
|
| 1110 |
+
{
|
| 1111 |
+
"epoch": 0.37,
|
| 1112 |
+
"learning_rate": 9.636067213442688e-05,
|
| 1113 |
+
"loss": 0.7912,
|
| 1114 |
+
"step": 18300
|
| 1115 |
+
},
|
| 1116 |
+
{
|
| 1117 |
+
"epoch": 0.37,
|
| 1118 |
+
"learning_rate": 9.634066813362673e-05,
|
| 1119 |
+
"loss": 0.7864,
|
| 1120 |
+
"step": 18400
|
| 1121 |
+
},
|
| 1122 |
+
{
|
| 1123 |
+
"epoch": 0.37,
|
| 1124 |
+
"learning_rate": 9.632066413282656e-05,
|
| 1125 |
+
"loss": 0.7639,
|
| 1126 |
+
"step": 18500
|
| 1127 |
+
},
|
| 1128 |
+
{
|
| 1129 |
+
"epoch": 0.37,
|
| 1130 |
+
"learning_rate": 9.63006601320264e-05,
|
| 1131 |
+
"loss": 0.7874,
|
| 1132 |
+
"step": 18600
|
| 1133 |
+
},
|
| 1134 |
+
{
|
| 1135 |
+
"epoch": 0.37,
|
| 1136 |
+
"learning_rate": 9.628065613122625e-05,
|
| 1137 |
+
"loss": 0.7642,
|
| 1138 |
+
"step": 18700
|
| 1139 |
+
},
|
| 1140 |
+
{
|
| 1141 |
+
"epoch": 0.38,
|
| 1142 |
+
"learning_rate": 9.626065213042609e-05,
|
| 1143 |
+
"loss": 0.756,
|
| 1144 |
+
"step": 18800
|
| 1145 |
+
},
|
| 1146 |
+
{
|
| 1147 |
+
"epoch": 0.38,
|
| 1148 |
+
"learning_rate": 9.624064812962594e-05,
|
| 1149 |
+
"loss": 0.8036,
|
| 1150 |
+
"step": 18900
|
| 1151 |
+
},
|
| 1152 |
+
{
|
| 1153 |
+
"epoch": 0.38,
|
| 1154 |
+
"learning_rate": 9.622064412882577e-05,
|
| 1155 |
+
"loss": 0.7537,
|
| 1156 |
+
"step": 19000
|
| 1157 |
+
},
|
| 1158 |
+
{
|
| 1159 |
+
"epoch": 0.38,
|
| 1160 |
+
"learning_rate": 9.620064012802561e-05,
|
| 1161 |
+
"loss": 0.8193,
|
| 1162 |
+
"step": 19100
|
| 1163 |
+
},
|
| 1164 |
+
{
|
| 1165 |
+
"epoch": 0.38,
|
| 1166 |
+
"learning_rate": 9.618063612722546e-05,
|
| 1167 |
+
"loss": 0.8036,
|
| 1168 |
+
"step": 19200
|
| 1169 |
+
},
|
| 1170 |
+
{
|
| 1171 |
+
"epoch": 0.39,
|
| 1172 |
+
"learning_rate": 9.616063212642529e-05,
|
| 1173 |
+
"loss": 0.784,
|
| 1174 |
+
"step": 19300
|
| 1175 |
+
},
|
| 1176 |
+
{
|
| 1177 |
+
"epoch": 0.39,
|
| 1178 |
+
"learning_rate": 9.614062812562513e-05,
|
| 1179 |
+
"loss": 0.7861,
|
| 1180 |
+
"step": 19400
|
| 1181 |
+
},
|
| 1182 |
+
{
|
| 1183 |
+
"epoch": 0.39,
|
| 1184 |
+
"learning_rate": 9.612062412482496e-05,
|
| 1185 |
+
"loss": 0.7882,
|
| 1186 |
+
"step": 19500
|
| 1187 |
+
},
|
| 1188 |
+
{
|
| 1189 |
+
"epoch": 0.39,
|
| 1190 |
+
"learning_rate": 9.610062012402481e-05,
|
| 1191 |
+
"loss": 0.8048,
|
| 1192 |
+
"step": 19600
|
| 1193 |
+
},
|
| 1194 |
+
{
|
| 1195 |
+
"epoch": 0.39,
|
| 1196 |
+
"learning_rate": 9.608061612322465e-05,
|
| 1197 |
+
"loss": 0.8544,
|
| 1198 |
+
"step": 19700
|
| 1199 |
+
},
|
| 1200 |
+
{
|
| 1201 |
+
"epoch": 0.4,
|
| 1202 |
+
"learning_rate": 9.606061212242448e-05,
|
| 1203 |
+
"loss": 0.7827,
|
| 1204 |
+
"step": 19800
|
| 1205 |
+
},
|
| 1206 |
+
{
|
| 1207 |
+
"epoch": 0.4,
|
| 1208 |
+
"learning_rate": 9.604060812162433e-05,
|
| 1209 |
+
"loss": 0.7641,
|
| 1210 |
+
"step": 19900
|
| 1211 |
+
},
|
| 1212 |
+
{
|
| 1213 |
+
"epoch": 0.4,
|
| 1214 |
+
"learning_rate": 9.602060412082416e-05,
|
| 1215 |
+
"loss": 0.7146,
|
| 1216 |
+
"step": 20000
|
| 1217 |
+
},
|
| 1218 |
+
{
|
| 1219 |
+
"epoch": 0.4,
|
| 1220 |
+
"eval_loss": 0.7286937832832336,
|
| 1221 |
+
"eval_runtime": 689.1627,
|
| 1222 |
+
"eval_samples_per_second": 7.217,
|
| 1223 |
+
"eval_steps_per_second": 0.903,
|
| 1224 |
+
"step": 20000
|
| 1225 |
+
},
|
| 1226 |
+
{
|
| 1227 |
+
"epoch": 0.4,
|
| 1228 |
+
"learning_rate": 9.6000600120024e-05,
|
| 1229 |
+
"loss": 0.7801,
|
| 1230 |
+
"step": 20100
|
| 1231 |
+
},
|
| 1232 |
+
{
|
| 1233 |
+
"epoch": 0.4,
|
| 1234 |
+
"learning_rate": 9.598059611922385e-05,
|
| 1235 |
+
"loss": 0.7803,
|
| 1236 |
+
"step": 20200
|
| 1237 |
+
},
|
| 1238 |
+
{
|
| 1239 |
+
"epoch": 0.41,
|
| 1240 |
+
"learning_rate": 9.596059211842369e-05,
|
| 1241 |
+
"loss": 0.7981,
|
| 1242 |
+
"step": 20300
|
| 1243 |
+
},
|
| 1244 |
+
{
|
| 1245 |
+
"epoch": 0.41,
|
| 1246 |
+
"learning_rate": 9.594058811762354e-05,
|
| 1247 |
+
"loss": 0.755,
|
| 1248 |
+
"step": 20400
|
| 1249 |
+
},
|
| 1250 |
+
{
|
| 1251 |
+
"epoch": 0.41,
|
| 1252 |
+
"learning_rate": 9.592058411682337e-05,
|
| 1253 |
+
"loss": 0.7743,
|
| 1254 |
+
"step": 20500
|
| 1255 |
+
},
|
| 1256 |
+
{
|
| 1257 |
+
"epoch": 0.41,
|
| 1258 |
+
"learning_rate": 9.590098019603922e-05,
|
| 1259 |
+
"loss": 0.7358,
|
| 1260 |
+
"step": 20600
|
| 1261 |
+
},
|
| 1262 |
+
{
|
| 1263 |
+
"epoch": 0.41,
|
| 1264 |
+
"learning_rate": 9.588097619523905e-05,
|
| 1265 |
+
"loss": 0.7851,
|
| 1266 |
+
"step": 20700
|
| 1267 |
+
},
|
| 1268 |
+
{
|
| 1269 |
+
"epoch": 0.42,
|
| 1270 |
+
"learning_rate": 9.586097219443889e-05,
|
| 1271 |
+
"loss": 0.7746,
|
| 1272 |
+
"step": 20800
|
| 1273 |
+
},
|
| 1274 |
+
{
|
| 1275 |
+
"epoch": 0.42,
|
| 1276 |
+
"learning_rate": 9.584096819363874e-05,
|
| 1277 |
+
"loss": 0.8246,
|
| 1278 |
+
"step": 20900
|
| 1279 |
+
},
|
| 1280 |
+
{
|
| 1281 |
+
"epoch": 0.42,
|
| 1282 |
+
"learning_rate": 9.582096419283857e-05,
|
| 1283 |
+
"loss": 0.7833,
|
| 1284 |
+
"step": 21000
|
| 1285 |
+
},
|
| 1286 |
+
{
|
| 1287 |
+
"epoch": 0.42,
|
| 1288 |
+
"learning_rate": 9.580096019203841e-05,
|
| 1289 |
+
"loss": 0.7666,
|
| 1290 |
+
"step": 21100
|
| 1291 |
+
},
|
| 1292 |
+
{
|
| 1293 |
+
"epoch": 0.42,
|
| 1294 |
+
"learning_rate": 9.578095619123824e-05,
|
| 1295 |
+
"loss": 0.8575,
|
| 1296 |
+
"step": 21200
|
| 1297 |
+
},
|
| 1298 |
+
{
|
| 1299 |
+
"epoch": 0.43,
|
| 1300 |
+
"learning_rate": 9.576095219043809e-05,
|
| 1301 |
+
"loss": 0.7934,
|
| 1302 |
+
"step": 21300
|
| 1303 |
+
},
|
| 1304 |
+
{
|
| 1305 |
+
"epoch": 0.43,
|
| 1306 |
+
"learning_rate": 9.574094818963793e-05,
|
| 1307 |
+
"loss": 0.7473,
|
| 1308 |
+
"step": 21400
|
| 1309 |
+
},
|
| 1310 |
+
{
|
| 1311 |
+
"epoch": 0.43,
|
| 1312 |
+
"learning_rate": 9.572094418883776e-05,
|
| 1313 |
+
"loss": 0.7455,
|
| 1314 |
+
"step": 21500
|
| 1315 |
+
},
|
| 1316 |
+
{
|
| 1317 |
+
"epoch": 0.43,
|
| 1318 |
+
"learning_rate": 9.570094018803761e-05,
|
| 1319 |
+
"loss": 0.7527,
|
| 1320 |
+
"step": 21600
|
| 1321 |
+
},
|
| 1322 |
+
{
|
| 1323 |
+
"epoch": 0.43,
|
| 1324 |
+
"learning_rate": 9.568093618723745e-05,
|
| 1325 |
+
"loss": 0.7906,
|
| 1326 |
+
"step": 21700
|
| 1327 |
+
},
|
| 1328 |
+
{
|
| 1329 |
+
"epoch": 0.44,
|
| 1330 |
+
"learning_rate": 9.56609321864373e-05,
|
| 1331 |
+
"loss": 0.8384,
|
| 1332 |
+
"step": 21800
|
| 1333 |
+
},
|
| 1334 |
+
{
|
| 1335 |
+
"epoch": 0.44,
|
| 1336 |
+
"learning_rate": 9.564092818563714e-05,
|
| 1337 |
+
"loss": 0.7897,
|
| 1338 |
+
"step": 21900
|
| 1339 |
+
},
|
| 1340 |
+
{
|
| 1341 |
+
"epoch": 0.44,
|
| 1342 |
+
"learning_rate": 9.562092418483697e-05,
|
| 1343 |
+
"loss": 0.7459,
|
| 1344 |
+
"step": 22000
|
| 1345 |
+
},
|
| 1346 |
+
{
|
| 1347 |
+
"epoch": 0.44,
|
| 1348 |
+
"learning_rate": 9.560092018403682e-05,
|
| 1349 |
+
"loss": 0.7838,
|
| 1350 |
+
"step": 22100
|
| 1351 |
+
},
|
| 1352 |
+
{
|
| 1353 |
+
"epoch": 0.44,
|
| 1354 |
+
"learning_rate": 9.558091618323666e-05,
|
| 1355 |
+
"loss": 0.7745,
|
| 1356 |
+
"step": 22200
|
| 1357 |
+
},
|
| 1358 |
+
{
|
| 1359 |
+
"epoch": 0.45,
|
| 1360 |
+
"learning_rate": 9.556091218243649e-05,
|
| 1361 |
+
"loss": 0.7985,
|
| 1362 |
+
"step": 22300
|
| 1363 |
+
},
|
| 1364 |
+
{
|
| 1365 |
+
"epoch": 0.45,
|
| 1366 |
+
"learning_rate": 9.554090818163634e-05,
|
| 1367 |
+
"loss": 0.7945,
|
| 1368 |
+
"step": 22400
|
| 1369 |
+
},
|
| 1370 |
+
{
|
| 1371 |
+
"epoch": 0.45,
|
| 1372 |
+
"learning_rate": 9.552090418083617e-05,
|
| 1373 |
+
"loss": 0.7985,
|
| 1374 |
+
"step": 22500
|
| 1375 |
+
},
|
| 1376 |
+
{
|
| 1377 |
+
"epoch": 0.45,
|
| 1378 |
+
"learning_rate": 9.550090018003601e-05,
|
| 1379 |
+
"loss": 0.7785,
|
| 1380 |
+
"step": 22600
|
| 1381 |
+
},
|
| 1382 |
+
{
|
| 1383 |
+
"epoch": 0.45,
|
| 1384 |
+
"learning_rate": 9.548089617923586e-05,
|
| 1385 |
+
"loss": 0.7642,
|
| 1386 |
+
"step": 22700
|
| 1387 |
+
},
|
| 1388 |
+
{
|
| 1389 |
+
"epoch": 0.46,
|
| 1390 |
+
"learning_rate": 9.546089217843569e-05,
|
| 1391 |
+
"loss": 0.7953,
|
| 1392 |
+
"step": 22800
|
| 1393 |
+
},
|
| 1394 |
+
{
|
| 1395 |
+
"epoch": 0.46,
|
| 1396 |
+
"learning_rate": 9.544088817763553e-05,
|
| 1397 |
+
"loss": 0.7848,
|
| 1398 |
+
"step": 22900
|
| 1399 |
+
},
|
| 1400 |
+
{
|
| 1401 |
+
"epoch": 0.46,
|
| 1402 |
+
"learning_rate": 9.542088417683536e-05,
|
| 1403 |
+
"loss": 0.7388,
|
| 1404 |
+
"step": 23000
|
| 1405 |
+
},
|
| 1406 |
+
{
|
| 1407 |
+
"epoch": 0.46,
|
| 1408 |
+
"learning_rate": 9.540088017603521e-05,
|
| 1409 |
+
"loss": 0.7947,
|
| 1410 |
+
"step": 23100
|
| 1411 |
+
},
|
| 1412 |
+
{
|
| 1413 |
+
"epoch": 0.46,
|
| 1414 |
+
"learning_rate": 9.538087617523505e-05,
|
| 1415 |
+
"loss": 0.7464,
|
| 1416 |
+
"step": 23200
|
| 1417 |
+
},
|
| 1418 |
+
{
|
| 1419 |
+
"epoch": 0.47,
|
| 1420 |
+
"learning_rate": 9.53608721744349e-05,
|
| 1421 |
+
"loss": 0.7696,
|
| 1422 |
+
"step": 23300
|
| 1423 |
+
},
|
| 1424 |
+
{
|
| 1425 |
+
"epoch": 0.47,
|
| 1426 |
+
"learning_rate": 9.534086817363474e-05,
|
| 1427 |
+
"loss": 0.7563,
|
| 1428 |
+
"step": 23400
|
| 1429 |
+
},
|
| 1430 |
+
{
|
| 1431 |
+
"epoch": 0.47,
|
| 1432 |
+
"learning_rate": 9.532086417283457e-05,
|
| 1433 |
+
"loss": 0.7522,
|
| 1434 |
+
"step": 23500
|
| 1435 |
+
},
|
| 1436 |
+
{
|
| 1437 |
+
"epoch": 0.47,
|
| 1438 |
+
"learning_rate": 9.530086017203442e-05,
|
| 1439 |
+
"loss": 0.772,
|
| 1440 |
+
"step": 23600
|
| 1441 |
+
},
|
| 1442 |
+
{
|
| 1443 |
+
"epoch": 0.47,
|
| 1444 |
+
"learning_rate": 9.528085617123426e-05,
|
| 1445 |
+
"loss": 0.7654,
|
| 1446 |
+
"step": 23700
|
| 1447 |
+
},
|
| 1448 |
+
{
|
| 1449 |
+
"epoch": 0.48,
|
| 1450 |
+
"learning_rate": 9.526085217043409e-05,
|
| 1451 |
+
"loss": 0.759,
|
| 1452 |
+
"step": 23800
|
| 1453 |
+
},
|
| 1454 |
+
{
|
| 1455 |
+
"epoch": 0.48,
|
| 1456 |
+
"learning_rate": 9.524084816963394e-05,
|
| 1457 |
+
"loss": 0.8032,
|
| 1458 |
+
"step": 23900
|
| 1459 |
+
},
|
| 1460 |
+
{
|
| 1461 |
+
"epoch": 0.48,
|
| 1462 |
+
"learning_rate": 9.522084416883377e-05,
|
| 1463 |
+
"loss": 0.7638,
|
| 1464 |
+
"step": 24000
|
| 1465 |
+
},
|
| 1466 |
+
{
|
| 1467 |
+
"epoch": 0.48,
|
| 1468 |
+
"learning_rate": 9.520084016803361e-05,
|
| 1469 |
+
"loss": 0.7974,
|
| 1470 |
+
"step": 24100
|
| 1471 |
+
},
|
| 1472 |
+
{
|
| 1473 |
+
"epoch": 0.48,
|
| 1474 |
+
"learning_rate": 9.518083616723346e-05,
|
| 1475 |
+
"loss": 0.7647,
|
| 1476 |
+
"step": 24200
|
| 1477 |
+
},
|
| 1478 |
+
{
|
| 1479 |
+
"epoch": 0.49,
|
| 1480 |
+
"learning_rate": 9.516083216643329e-05,
|
| 1481 |
+
"loss": 0.8016,
|
| 1482 |
+
"step": 24300
|
| 1483 |
+
},
|
| 1484 |
+
{
|
| 1485 |
+
"epoch": 0.49,
|
| 1486 |
+
"learning_rate": 9.514082816563313e-05,
|
| 1487 |
+
"loss": 0.7749,
|
| 1488 |
+
"step": 24400
|
| 1489 |
+
},
|
| 1490 |
+
{
|
| 1491 |
+
"epoch": 0.49,
|
| 1492 |
+
"learning_rate": 9.512082416483296e-05,
|
| 1493 |
+
"loss": 0.7573,
|
| 1494 |
+
"step": 24500
|
| 1495 |
+
},
|
| 1496 |
+
{
|
| 1497 |
+
"epoch": 0.49,
|
| 1498 |
+
"learning_rate": 9.51008201640328e-05,
|
| 1499 |
+
"loss": 0.7476,
|
| 1500 |
+
"step": 24600
|
| 1501 |
+
},
|
| 1502 |
+
{
|
| 1503 |
+
"epoch": 0.49,
|
| 1504 |
+
"learning_rate": 9.508081616323265e-05,
|
| 1505 |
+
"loss": 0.7848,
|
| 1506 |
+
"step": 24700
|
| 1507 |
+
},
|
| 1508 |
+
{
|
| 1509 |
+
"epoch": 0.5,
|
| 1510 |
+
"learning_rate": 9.50608121624325e-05,
|
| 1511 |
+
"loss": 0.7994,
|
| 1512 |
+
"step": 24800
|
| 1513 |
+
},
|
| 1514 |
+
{
|
| 1515 |
+
"epoch": 0.5,
|
| 1516 |
+
"learning_rate": 9.504100820164034e-05,
|
| 1517 |
+
"loss": 0.8056,
|
| 1518 |
+
"step": 24900
|
| 1519 |
+
},
|
| 1520 |
+
{
|
| 1521 |
+
"epoch": 0.5,
|
| 1522 |
+
"learning_rate": 9.502100420084017e-05,
|
| 1523 |
+
"loss": 0.8066,
|
| 1524 |
+
"step": 25000
|
| 1525 |
+
},
|
| 1526 |
+
{
|
| 1527 |
+
"epoch": 0.5,
|
| 1528 |
+
"learning_rate": 9.500100020004001e-05,
|
| 1529 |
+
"loss": 0.7506,
|
| 1530 |
+
"step": 25100
|
| 1531 |
+
},
|
| 1532 |
+
{
|
| 1533 |
+
"epoch": 0.5,
|
| 1534 |
+
"learning_rate": 9.498099619923986e-05,
|
| 1535 |
+
"loss": 0.8018,
|
| 1536 |
+
"step": 25200
|
| 1537 |
+
},
|
| 1538 |
+
{
|
| 1539 |
+
"epoch": 0.51,
|
| 1540 |
+
"learning_rate": 9.496099219843969e-05,
|
| 1541 |
+
"loss": 0.7451,
|
| 1542 |
+
"step": 25300
|
| 1543 |
+
},
|
| 1544 |
+
{
|
| 1545 |
+
"epoch": 0.51,
|
| 1546 |
+
"learning_rate": 9.494098819763953e-05,
|
| 1547 |
+
"loss": 0.764,
|
| 1548 |
+
"step": 25400
|
| 1549 |
+
},
|
| 1550 |
+
{
|
| 1551 |
+
"epoch": 0.51,
|
| 1552 |
+
"learning_rate": 9.492098419683938e-05,
|
| 1553 |
+
"loss": 0.8212,
|
| 1554 |
+
"step": 25500
|
| 1555 |
+
},
|
| 1556 |
+
{
|
| 1557 |
+
"epoch": 0.51,
|
| 1558 |
+
"learning_rate": 9.490098019603922e-05,
|
| 1559 |
+
"loss": 0.7954,
|
| 1560 |
+
"step": 25600
|
| 1561 |
+
},
|
| 1562 |
+
{
|
| 1563 |
+
"epoch": 0.51,
|
| 1564 |
+
"learning_rate": 9.488097619523905e-05,
|
| 1565 |
+
"loss": 0.7965,
|
| 1566 |
+
"step": 25700
|
| 1567 |
+
},
|
| 1568 |
+
{
|
| 1569 |
+
"epoch": 0.52,
|
| 1570 |
+
"learning_rate": 9.48609721944389e-05,
|
| 1571 |
+
"loss": 0.7795,
|
| 1572 |
+
"step": 25800
|
| 1573 |
+
},
|
| 1574 |
+
{
|
| 1575 |
+
"epoch": 0.52,
|
| 1576 |
+
"learning_rate": 9.484096819363874e-05,
|
| 1577 |
+
"loss": 0.7993,
|
| 1578 |
+
"step": 25900
|
| 1579 |
+
},
|
| 1580 |
+
{
|
| 1581 |
+
"epoch": 0.52,
|
| 1582 |
+
"learning_rate": 9.482096419283857e-05,
|
| 1583 |
+
"loss": 0.7951,
|
| 1584 |
+
"step": 26000
|
| 1585 |
+
},
|
| 1586 |
+
{
|
| 1587 |
+
"epoch": 0.52,
|
| 1588 |
+
"learning_rate": 9.480096019203842e-05,
|
| 1589 |
+
"loss": 0.7432,
|
| 1590 |
+
"step": 26100
|
| 1591 |
+
},
|
| 1592 |
+
{
|
| 1593 |
+
"epoch": 0.52,
|
| 1594 |
+
"learning_rate": 9.478095619123826e-05,
|
| 1595 |
+
"loss": 0.7844,
|
| 1596 |
+
"step": 26200
|
| 1597 |
+
},
|
| 1598 |
+
{
|
| 1599 |
+
"epoch": 0.53,
|
| 1600 |
+
"learning_rate": 9.476095219043809e-05,
|
| 1601 |
+
"loss": 0.7029,
|
| 1602 |
+
"step": 26300
|
| 1603 |
+
},
|
| 1604 |
+
{
|
| 1605 |
+
"epoch": 0.53,
|
| 1606 |
+
"learning_rate": 9.474094818963793e-05,
|
| 1607 |
+
"loss": 0.7613,
|
| 1608 |
+
"step": 26400
|
| 1609 |
+
},
|
| 1610 |
+
{
|
| 1611 |
+
"epoch": 0.53,
|
| 1612 |
+
"learning_rate": 9.472094418883777e-05,
|
| 1613 |
+
"loss": 0.7896,
|
| 1614 |
+
"step": 26500
|
| 1615 |
+
},
|
| 1616 |
+
{
|
| 1617 |
+
"epoch": 0.53,
|
| 1618 |
+
"learning_rate": 9.470094018803761e-05,
|
| 1619 |
+
"loss": 0.7674,
|
| 1620 |
+
"step": 26600
|
| 1621 |
+
},
|
| 1622 |
+
{
|
| 1623 |
+
"epoch": 0.53,
|
| 1624 |
+
"learning_rate": 9.468093618723745e-05,
|
| 1625 |
+
"loss": 0.8013,
|
| 1626 |
+
"step": 26700
|
| 1627 |
+
},
|
| 1628 |
+
{
|
| 1629 |
+
"epoch": 0.54,
|
| 1630 |
+
"learning_rate": 9.466093218643729e-05,
|
| 1631 |
+
"loss": 0.811,
|
| 1632 |
+
"step": 26800
|
| 1633 |
+
},
|
| 1634 |
+
{
|
| 1635 |
+
"epoch": 0.54,
|
| 1636 |
+
"learning_rate": 9.464092818563713e-05,
|
| 1637 |
+
"loss": 0.8235,
|
| 1638 |
+
"step": 26900
|
| 1639 |
+
},
|
| 1640 |
+
{
|
| 1641 |
+
"epoch": 0.54,
|
| 1642 |
+
"learning_rate": 9.462092418483696e-05,
|
| 1643 |
+
"loss": 0.7717,
|
| 1644 |
+
"step": 27000
|
| 1645 |
+
},
|
| 1646 |
+
{
|
| 1647 |
+
"epoch": 0.54,
|
| 1648 |
+
"learning_rate": 9.46009201840368e-05,
|
| 1649 |
+
"loss": 0.7364,
|
| 1650 |
+
"step": 27100
|
| 1651 |
+
},
|
| 1652 |
+
{
|
| 1653 |
+
"epoch": 0.54,
|
| 1654 |
+
"learning_rate": 9.458111622324465e-05,
|
| 1655 |
+
"loss": 0.7663,
|
| 1656 |
+
"step": 27200
|
| 1657 |
+
},
|
| 1658 |
+
{
|
| 1659 |
+
"epoch": 0.55,
|
| 1660 |
+
"learning_rate": 9.456111222244449e-05,
|
| 1661 |
+
"loss": 0.771,
|
| 1662 |
+
"step": 27300
|
| 1663 |
+
},
|
| 1664 |
+
{
|
| 1665 |
+
"epoch": 0.55,
|
| 1666 |
+
"learning_rate": 9.454110822164433e-05,
|
| 1667 |
+
"loss": 0.7583,
|
| 1668 |
+
"step": 27400
|
| 1669 |
+
},
|
| 1670 |
+
{
|
| 1671 |
+
"epoch": 0.55,
|
| 1672 |
+
"learning_rate": 9.452110422084417e-05,
|
| 1673 |
+
"loss": 0.7672,
|
| 1674 |
+
"step": 27500
|
| 1675 |
+
},
|
| 1676 |
+
{
|
| 1677 |
+
"epoch": 0.55,
|
| 1678 |
+
"learning_rate": 9.450110022004401e-05,
|
| 1679 |
+
"loss": 0.7706,
|
| 1680 |
+
"step": 27600
|
| 1681 |
+
},
|
| 1682 |
+
{
|
| 1683 |
+
"epoch": 0.55,
|
| 1684 |
+
"learning_rate": 9.448109621924385e-05,
|
| 1685 |
+
"loss": 0.7676,
|
| 1686 |
+
"step": 27700
|
| 1687 |
+
},
|
| 1688 |
+
{
|
| 1689 |
+
"epoch": 0.56,
|
| 1690 |
+
"learning_rate": 9.446109221844369e-05,
|
| 1691 |
+
"loss": 0.7859,
|
| 1692 |
+
"step": 27800
|
| 1693 |
+
},
|
| 1694 |
+
{
|
| 1695 |
+
"epoch": 0.56,
|
| 1696 |
+
"learning_rate": 9.444108821764353e-05,
|
| 1697 |
+
"loss": 0.7573,
|
| 1698 |
+
"step": 27900
|
| 1699 |
+
},
|
| 1700 |
+
{
|
| 1701 |
+
"epoch": 0.56,
|
| 1702 |
+
"learning_rate": 9.442108421684337e-05,
|
| 1703 |
+
"loss": 0.8289,
|
| 1704 |
+
"step": 28000
|
| 1705 |
+
},
|
| 1706 |
+
{
|
| 1707 |
+
"epoch": 0.56,
|
| 1708 |
+
"learning_rate": 9.440108021604322e-05,
|
| 1709 |
+
"loss": 0.7757,
|
| 1710 |
+
"step": 28100
|
| 1711 |
+
},
|
| 1712 |
+
{
|
| 1713 |
+
"epoch": 0.56,
|
| 1714 |
+
"learning_rate": 9.438107621524306e-05,
|
| 1715 |
+
"loss": 0.7688,
|
| 1716 |
+
"step": 28200
|
| 1717 |
+
},
|
| 1718 |
+
{
|
| 1719 |
+
"epoch": 0.57,
|
| 1720 |
+
"learning_rate": 9.43610722144429e-05,
|
| 1721 |
+
"loss": 0.7941,
|
| 1722 |
+
"step": 28300
|
| 1723 |
+
},
|
| 1724 |
+
{
|
| 1725 |
+
"epoch": 0.57,
|
| 1726 |
+
"learning_rate": 9.434106821364274e-05,
|
| 1727 |
+
"loss": 0.7829,
|
| 1728 |
+
"step": 28400
|
| 1729 |
+
},
|
| 1730 |
+
{
|
| 1731 |
+
"epoch": 0.57,
|
| 1732 |
+
"learning_rate": 9.432106421284257e-05,
|
| 1733 |
+
"loss": 0.7744,
|
| 1734 |
+
"step": 28500
|
| 1735 |
+
},
|
| 1736 |
+
{
|
| 1737 |
+
"epoch": 0.57,
|
| 1738 |
+
"learning_rate": 9.430106021204241e-05,
|
| 1739 |
+
"loss": 0.805,
|
| 1740 |
+
"step": 28600
|
| 1741 |
+
},
|
| 1742 |
+
{
|
| 1743 |
+
"epoch": 0.57,
|
| 1744 |
+
"learning_rate": 9.428105621124226e-05,
|
| 1745 |
+
"loss": 0.776,
|
| 1746 |
+
"step": 28700
|
| 1747 |
+
},
|
| 1748 |
+
{
|
| 1749 |
+
"epoch": 0.58,
|
| 1750 |
+
"learning_rate": 9.426105221044209e-05,
|
| 1751 |
+
"loss": 0.7839,
|
| 1752 |
+
"step": 28800
|
| 1753 |
+
},
|
| 1754 |
+
{
|
| 1755 |
+
"epoch": 0.58,
|
| 1756 |
+
"learning_rate": 9.424104820964193e-05,
|
| 1757 |
+
"loss": 0.7894,
|
| 1758 |
+
"step": 28900
|
| 1759 |
+
},
|
| 1760 |
+
{
|
| 1761 |
+
"epoch": 0.58,
|
| 1762 |
+
"learning_rate": 9.422104420884177e-05,
|
| 1763 |
+
"loss": 0.8138,
|
| 1764 |
+
"step": 29000
|
| 1765 |
+
},
|
| 1766 |
+
{
|
| 1767 |
+
"epoch": 0.58,
|
| 1768 |
+
"learning_rate": 9.420104020804161e-05,
|
| 1769 |
+
"loss": 0.7834,
|
| 1770 |
+
"step": 29100
|
| 1771 |
+
},
|
| 1772 |
+
{
|
| 1773 |
+
"epoch": 0.58,
|
| 1774 |
+
"learning_rate": 9.418103620724145e-05,
|
| 1775 |
+
"loss": 0.7323,
|
| 1776 |
+
"step": 29200
|
| 1777 |
+
},
|
| 1778 |
+
{
|
| 1779 |
+
"epoch": 0.59,
|
| 1780 |
+
"learning_rate": 9.416103220644129e-05,
|
| 1781 |
+
"loss": 0.79,
|
| 1782 |
+
"step": 29300
|
| 1783 |
+
},
|
| 1784 |
+
{
|
| 1785 |
+
"epoch": 0.59,
|
| 1786 |
+
"learning_rate": 9.414102820564113e-05,
|
| 1787 |
+
"loss": 0.773,
|
| 1788 |
+
"step": 29400
|
| 1789 |
+
},
|
| 1790 |
+
{
|
| 1791 |
+
"epoch": 0.59,
|
| 1792 |
+
"learning_rate": 9.412102420484097e-05,
|
| 1793 |
+
"loss": 0.7607,
|
| 1794 |
+
"step": 29500
|
| 1795 |
+
},
|
| 1796 |
+
{
|
| 1797 |
+
"epoch": 0.59,
|
| 1798 |
+
"learning_rate": 9.410102020404082e-05,
|
| 1799 |
+
"loss": 0.7581,
|
| 1800 |
+
"step": 29600
|
| 1801 |
+
},
|
| 1802 |
+
{
|
| 1803 |
+
"epoch": 0.59,
|
| 1804 |
+
"learning_rate": 9.408101620324066e-05,
|
| 1805 |
+
"loss": 0.7528,
|
| 1806 |
+
"step": 29700
|
| 1807 |
+
},
|
| 1808 |
+
{
|
| 1809 |
+
"epoch": 0.6,
|
| 1810 |
+
"learning_rate": 9.40610122024405e-05,
|
| 1811 |
+
"loss": 0.7755,
|
| 1812 |
+
"step": 29800
|
| 1813 |
+
},
|
| 1814 |
+
{
|
| 1815 |
+
"epoch": 0.6,
|
| 1816 |
+
"learning_rate": 9.404100820164034e-05,
|
| 1817 |
+
"loss": 0.7633,
|
| 1818 |
+
"step": 29900
|
| 1819 |
+
},
|
| 1820 |
+
{
|
| 1821 |
+
"epoch": 0.6,
|
| 1822 |
+
"learning_rate": 9.402100420084017e-05,
|
| 1823 |
+
"loss": 0.7352,
|
| 1824 |
+
"step": 30000
|
| 1825 |
+
},
|
| 1826 |
+
{
|
| 1827 |
+
"epoch": 0.6,
|
| 1828 |
+
"eval_loss": 0.7202807068824768,
|
| 1829 |
+
"eval_runtime": 707.6198,
|
| 1830 |
+
"eval_samples_per_second": 7.029,
|
| 1831 |
+
"eval_steps_per_second": 0.879,
|
| 1832 |
+
"step": 30000
|
| 1833 |
+
}
|
| 1834 |
+
],
|
| 1835 |
+
"max_steps": 500000,
|
| 1836 |
+
"num_train_epochs": 10,
|
| 1837 |
+
"total_flos": 4.209735834528645e+17,
|
| 1838 |
+
"trial_name": null,
|
| 1839 |
+
"trial_params": null
|
| 1840 |
+
}
|
checkpoint-30000/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:86e3e950534e1e3e08a50aef6d282bea64ab70d530ec27d5c1bcac285ba164ae
|
| 3 |
+
size 4408
|
checkpoint-40000/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<pad>": 32000
|
| 3 |
+
}
|
checkpoint-40000/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6badfad5ecc03e9e8fd354c02e1561d20668442b7f4b25b038c1650f03d2fdc5
|
| 3 |
+
size 134433530
|
checkpoint-40000/pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:60b57d5e5f1f21becd36117caae3f5893aa205ff0fc43eada70276984740c4e1
|
| 3 |
+
size 67201802
|
checkpoint-40000/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:21bd6a88258929b4bf272f90af5288e75f1da0c2897f1b6dca050e98db489c00
|
| 3 |
+
size 14244
|
checkpoint-40000/scaler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e310416331bcfbb2f7b543742f04dddf827fd8474acc16e4758f26e1f2bc21d6
|
| 3 |
+
size 988
|
checkpoint-40000/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:621dde525b20ac0ab00890dd6ac4c39c5e6172380975e17232852e8210b1b7cc
|
| 3 |
+
size 1064
|
checkpoint-40000/special_tokens_map.json
ADDED
|
@@ -0,0 +1,24 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"content": "<s>",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": false,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"eos_token": {
|
| 10 |
+
"content": "</s>",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"pad_token": "<s>",
|
| 17 |
+
"unk_token": {
|
| 18 |
+
"content": "<unk>",
|
| 19 |
+
"lstrip": false,
|
| 20 |
+
"normalized": false,
|
| 21 |
+
"rstrip": false,
|
| 22 |
+
"single_word": false
|
| 23 |
+
}
|
| 24 |
+
}
|
checkpoint-40000/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
checkpoint-40000/tokenizer.model
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9e556afd44213b6bd1be2b850ebbbd98f5481437a8021afaf58ee7fb1818d347
|
| 3 |
+
size 499723
|
checkpoint-40000/tokenizer_config.json
ADDED
|
@@ -0,0 +1,34 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_bos_token": true,
|
| 3 |
+
"add_eos_token": false,
|
| 4 |
+
"bos_token": {
|
| 5 |
+
"__type": "AddedToken",
|
| 6 |
+
"content": "<s>",
|
| 7 |
+
"lstrip": false,
|
| 8 |
+
"normalized": false,
|
| 9 |
+
"rstrip": false,
|
| 10 |
+
"single_word": false
|
| 11 |
+
},
|
| 12 |
+
"clean_up_tokenization_spaces": false,
|
| 13 |
+
"eos_token": {
|
| 14 |
+
"__type": "AddedToken",
|
| 15 |
+
"content": "</s>",
|
| 16 |
+
"lstrip": false,
|
| 17 |
+
"normalized": false,
|
| 18 |
+
"rstrip": false,
|
| 19 |
+
"single_word": false
|
| 20 |
+
},
|
| 21 |
+
"legacy": false,
|
| 22 |
+
"model_max_length": 1000000000000000019884624838656,
|
| 23 |
+
"pad_token": null,
|
| 24 |
+
"sp_model_kwargs": {},
|
| 25 |
+
"tokenizer_class": "LlamaTokenizer",
|
| 26 |
+
"unk_token": {
|
| 27 |
+
"__type": "AddedToken",
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false
|
| 33 |
+
}
|
| 34 |
+
}
|
checkpoint-40000/trainer_state.json
ADDED
|
@@ -0,0 +1,2448 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.7169719338417053,
|
| 3 |
+
"best_model_checkpoint": "/media/xzuyn/c41d93ed-83a7-4577-9d38-5062b7e14438/xzuyn/Documents/AI_ROCM_5-6/LLaMa-LoRA-Tuner/data/lora_models/LLaMa-2-WizardLM-Uncensored-Tulu-Format-50K-7B-LoRA/checkpoint-40000",
|
| 4 |
+
"epoch": 0.8,
|
| 5 |
+
"global_step": 40000,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.0,
|
| 12 |
+
"learning_rate": 0.0001,
|
| 13 |
+
"loss": 1.1136,
|
| 14 |
+
"step": 100
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.0,
|
| 18 |
+
"learning_rate": 9.997999599919985e-05,
|
| 19 |
+
"loss": 0.9595,
|
| 20 |
+
"step": 200
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.01,
|
| 24 |
+
"learning_rate": 9.995999199839968e-05,
|
| 25 |
+
"loss": 0.8936,
|
| 26 |
+
"step": 300
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.01,
|
| 30 |
+
"learning_rate": 9.993998799759952e-05,
|
| 31 |
+
"loss": 0.9237,
|
| 32 |
+
"step": 400
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.01,
|
| 36 |
+
"learning_rate": 9.992038407681536e-05,
|
| 37 |
+
"loss": 0.9005,
|
| 38 |
+
"step": 500
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.01,
|
| 42 |
+
"learning_rate": 9.99003800760152e-05,
|
| 43 |
+
"loss": 0.8433,
|
| 44 |
+
"step": 600
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.01,
|
| 48 |
+
"learning_rate": 9.988037607521505e-05,
|
| 49 |
+
"loss": 0.863,
|
| 50 |
+
"step": 700
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.02,
|
| 54 |
+
"learning_rate": 9.98603720744149e-05,
|
| 55 |
+
"loss": 0.8634,
|
| 56 |
+
"step": 800
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.02,
|
| 60 |
+
"learning_rate": 9.984036807361474e-05,
|
| 61 |
+
"loss": 0.8463,
|
| 62 |
+
"step": 900
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.02,
|
| 66 |
+
"learning_rate": 9.982036407281457e-05,
|
| 67 |
+
"loss": 0.8527,
|
| 68 |
+
"step": 1000
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.02,
|
| 72 |
+
"learning_rate": 9.980036007201441e-05,
|
| 73 |
+
"loss": 0.8259,
|
| 74 |
+
"step": 1100
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.02,
|
| 78 |
+
"learning_rate": 9.978035607121424e-05,
|
| 79 |
+
"loss": 0.8473,
|
| 80 |
+
"step": 1200
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.03,
|
| 84 |
+
"learning_rate": 9.976035207041409e-05,
|
| 85 |
+
"loss": 0.8462,
|
| 86 |
+
"step": 1300
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.03,
|
| 90 |
+
"learning_rate": 9.974034806961393e-05,
|
| 91 |
+
"loss": 0.8609,
|
| 92 |
+
"step": 1400
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.03,
|
| 96 |
+
"learning_rate": 9.972034406881376e-05,
|
| 97 |
+
"loss": 0.8238,
|
| 98 |
+
"step": 1500
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.03,
|
| 102 |
+
"learning_rate": 9.970034006801361e-05,
|
| 103 |
+
"loss": 0.8212,
|
| 104 |
+
"step": 1600
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 0.03,
|
| 108 |
+
"learning_rate": 9.968033606721344e-05,
|
| 109 |
+
"loss": 0.8391,
|
| 110 |
+
"step": 1700
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.04,
|
| 114 |
+
"learning_rate": 9.966033206641328e-05,
|
| 115 |
+
"loss": 0.774,
|
| 116 |
+
"step": 1800
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 0.04,
|
| 120 |
+
"learning_rate": 9.964032806561313e-05,
|
| 121 |
+
"loss": 0.8149,
|
| 122 |
+
"step": 1900
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.04,
|
| 126 |
+
"learning_rate": 9.962032406481296e-05,
|
| 127 |
+
"loss": 0.8794,
|
| 128 |
+
"step": 2000
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.04,
|
| 132 |
+
"learning_rate": 9.96003200640128e-05,
|
| 133 |
+
"loss": 0.7747,
|
| 134 |
+
"step": 2100
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.04,
|
| 138 |
+
"learning_rate": 9.958031606321264e-05,
|
| 139 |
+
"loss": 0.8191,
|
| 140 |
+
"step": 2200
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 0.05,
|
| 144 |
+
"learning_rate": 9.956031206241248e-05,
|
| 145 |
+
"loss": 0.8219,
|
| 146 |
+
"step": 2300
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 0.05,
|
| 150 |
+
"learning_rate": 9.954030806161232e-05,
|
| 151 |
+
"loss": 0.8191,
|
| 152 |
+
"step": 2400
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 0.05,
|
| 156 |
+
"learning_rate": 9.952030406081217e-05,
|
| 157 |
+
"loss": 0.8662,
|
| 158 |
+
"step": 2500
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 0.05,
|
| 162 |
+
"learning_rate": 9.950030006001201e-05,
|
| 163 |
+
"loss": 0.787,
|
| 164 |
+
"step": 2600
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.05,
|
| 168 |
+
"learning_rate": 9.948029605921184e-05,
|
| 169 |
+
"loss": 0.8668,
|
| 170 |
+
"step": 2700
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.06,
|
| 174 |
+
"learning_rate": 9.946029205841169e-05,
|
| 175 |
+
"loss": 0.7817,
|
| 176 |
+
"step": 2800
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 0.06,
|
| 180 |
+
"learning_rate": 9.944028805761153e-05,
|
| 181 |
+
"loss": 0.8021,
|
| 182 |
+
"step": 2900
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 0.06,
|
| 186 |
+
"learning_rate": 9.942028405681136e-05,
|
| 187 |
+
"loss": 0.8324,
|
| 188 |
+
"step": 3000
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 0.06,
|
| 192 |
+
"learning_rate": 9.940028005601121e-05,
|
| 193 |
+
"loss": 0.8631,
|
| 194 |
+
"step": 3100
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 0.06,
|
| 198 |
+
"learning_rate": 9.938027605521105e-05,
|
| 199 |
+
"loss": 0.797,
|
| 200 |
+
"step": 3200
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 0.07,
|
| 204 |
+
"learning_rate": 9.936027205441088e-05,
|
| 205 |
+
"loss": 0.8224,
|
| 206 |
+
"step": 3300
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.07,
|
| 210 |
+
"learning_rate": 9.934026805361073e-05,
|
| 211 |
+
"loss": 0.8164,
|
| 212 |
+
"step": 3400
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.07,
|
| 216 |
+
"learning_rate": 9.932026405281056e-05,
|
| 217 |
+
"loss": 0.7975,
|
| 218 |
+
"step": 3500
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 0.07,
|
| 222 |
+
"learning_rate": 9.930046009201841e-05,
|
| 223 |
+
"loss": 0.8468,
|
| 224 |
+
"step": 3600
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 0.07,
|
| 228 |
+
"learning_rate": 9.928045609121824e-05,
|
| 229 |
+
"loss": 0.8169,
|
| 230 |
+
"step": 3700
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 0.08,
|
| 234 |
+
"learning_rate": 9.926045209041809e-05,
|
| 235 |
+
"loss": 0.7761,
|
| 236 |
+
"step": 3800
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 0.08,
|
| 240 |
+
"learning_rate": 9.924044808961793e-05,
|
| 241 |
+
"loss": 0.8042,
|
| 242 |
+
"step": 3900
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 0.08,
|
| 246 |
+
"learning_rate": 9.922044408881776e-05,
|
| 247 |
+
"loss": 0.8146,
|
| 248 |
+
"step": 4000
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 0.08,
|
| 252 |
+
"learning_rate": 9.920044008801761e-05,
|
| 253 |
+
"loss": 0.8243,
|
| 254 |
+
"step": 4100
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.08,
|
| 258 |
+
"learning_rate": 9.918043608721744e-05,
|
| 259 |
+
"loss": 0.8315,
|
| 260 |
+
"step": 4200
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 0.09,
|
| 264 |
+
"learning_rate": 9.916043208641728e-05,
|
| 265 |
+
"loss": 0.8014,
|
| 266 |
+
"step": 4300
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 0.09,
|
| 270 |
+
"learning_rate": 9.914042808561713e-05,
|
| 271 |
+
"loss": 0.8241,
|
| 272 |
+
"step": 4400
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 0.09,
|
| 276 |
+
"learning_rate": 9.912042408481696e-05,
|
| 277 |
+
"loss": 0.8233,
|
| 278 |
+
"step": 4500
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 0.09,
|
| 282 |
+
"learning_rate": 9.91004200840168e-05,
|
| 283 |
+
"loss": 0.7748,
|
| 284 |
+
"step": 4600
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 0.09,
|
| 288 |
+
"learning_rate": 9.908041608321665e-05,
|
| 289 |
+
"loss": 0.8197,
|
| 290 |
+
"step": 4700
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 0.1,
|
| 294 |
+
"learning_rate": 9.906041208241649e-05,
|
| 295 |
+
"loss": 0.8195,
|
| 296 |
+
"step": 4800
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.1,
|
| 300 |
+
"learning_rate": 9.904040808161634e-05,
|
| 301 |
+
"loss": 0.7864,
|
| 302 |
+
"step": 4900
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 0.1,
|
| 306 |
+
"learning_rate": 9.902040408081617e-05,
|
| 307 |
+
"loss": 0.7812,
|
| 308 |
+
"step": 5000
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 0.1,
|
| 312 |
+
"learning_rate": 9.900040008001601e-05,
|
| 313 |
+
"loss": 0.7966,
|
| 314 |
+
"step": 5100
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 0.1,
|
| 318 |
+
"learning_rate": 9.898039607921584e-05,
|
| 319 |
+
"loss": 0.793,
|
| 320 |
+
"step": 5200
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 0.11,
|
| 324 |
+
"learning_rate": 9.896039207841569e-05,
|
| 325 |
+
"loss": 0.8568,
|
| 326 |
+
"step": 5300
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 0.11,
|
| 330 |
+
"learning_rate": 9.894038807761553e-05,
|
| 331 |
+
"loss": 0.7614,
|
| 332 |
+
"step": 5400
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 0.11,
|
| 336 |
+
"learning_rate": 9.892038407681536e-05,
|
| 337 |
+
"loss": 0.8171,
|
| 338 |
+
"step": 5500
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.11,
|
| 342 |
+
"learning_rate": 9.890038007601521e-05,
|
| 343 |
+
"loss": 0.7649,
|
| 344 |
+
"step": 5600
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 0.11,
|
| 348 |
+
"learning_rate": 9.888037607521505e-05,
|
| 349 |
+
"loss": 0.8074,
|
| 350 |
+
"step": 5700
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 0.12,
|
| 354 |
+
"learning_rate": 9.886037207441488e-05,
|
| 355 |
+
"loss": 0.7591,
|
| 356 |
+
"step": 5800
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 0.12,
|
| 360 |
+
"learning_rate": 9.884036807361473e-05,
|
| 361 |
+
"loss": 0.8102,
|
| 362 |
+
"step": 5900
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 0.12,
|
| 366 |
+
"learning_rate": 9.882036407281456e-05,
|
| 367 |
+
"loss": 0.784,
|
| 368 |
+
"step": 6000
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 0.12,
|
| 372 |
+
"learning_rate": 9.88003600720144e-05,
|
| 373 |
+
"loss": 0.8367,
|
| 374 |
+
"step": 6100
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 0.12,
|
| 378 |
+
"learning_rate": 9.878035607121425e-05,
|
| 379 |
+
"loss": 0.8418,
|
| 380 |
+
"step": 6200
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.13,
|
| 384 |
+
"learning_rate": 9.876035207041409e-05,
|
| 385 |
+
"loss": 0.798,
|
| 386 |
+
"step": 6300
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 0.13,
|
| 390 |
+
"learning_rate": 9.874034806961394e-05,
|
| 391 |
+
"loss": 0.7928,
|
| 392 |
+
"step": 6400
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 0.13,
|
| 396 |
+
"learning_rate": 9.872034406881377e-05,
|
| 397 |
+
"loss": 0.8262,
|
| 398 |
+
"step": 6500
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 0.13,
|
| 402 |
+
"learning_rate": 9.870034006801361e-05,
|
| 403 |
+
"loss": 0.7965,
|
| 404 |
+
"step": 6600
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 0.13,
|
| 408 |
+
"learning_rate": 9.868033606721346e-05,
|
| 409 |
+
"loss": 0.7952,
|
| 410 |
+
"step": 6700
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 0.14,
|
| 414 |
+
"learning_rate": 9.866033206641329e-05,
|
| 415 |
+
"loss": 0.7987,
|
| 416 |
+
"step": 6800
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 0.14,
|
| 420 |
+
"learning_rate": 9.864032806561313e-05,
|
| 421 |
+
"loss": 0.7939,
|
| 422 |
+
"step": 6900
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.14,
|
| 426 |
+
"learning_rate": 9.862032406481296e-05,
|
| 427 |
+
"loss": 0.7484,
|
| 428 |
+
"step": 7000
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 0.14,
|
| 432 |
+
"learning_rate": 9.860032006401281e-05,
|
| 433 |
+
"loss": 0.819,
|
| 434 |
+
"step": 7100
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 0.14,
|
| 438 |
+
"learning_rate": 9.858031606321265e-05,
|
| 439 |
+
"loss": 0.747,
|
| 440 |
+
"step": 7200
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 0.15,
|
| 444 |
+
"learning_rate": 9.856031206241248e-05,
|
| 445 |
+
"loss": 0.8029,
|
| 446 |
+
"step": 7300
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 0.15,
|
| 450 |
+
"learning_rate": 9.854030806161233e-05,
|
| 451 |
+
"loss": 0.9013,
|
| 452 |
+
"step": 7400
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 0.15,
|
| 456 |
+
"learning_rate": 9.852030406081216e-05,
|
| 457 |
+
"loss": 0.8337,
|
| 458 |
+
"step": 7500
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 0.15,
|
| 462 |
+
"learning_rate": 9.8500300060012e-05,
|
| 463 |
+
"loss": 0.7927,
|
| 464 |
+
"step": 7600
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.15,
|
| 468 |
+
"learning_rate": 9.848029605921185e-05,
|
| 469 |
+
"loss": 0.8725,
|
| 470 |
+
"step": 7700
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 0.16,
|
| 474 |
+
"learning_rate": 9.846029205841169e-05,
|
| 475 |
+
"loss": 0.7865,
|
| 476 |
+
"step": 7800
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 0.16,
|
| 480 |
+
"learning_rate": 9.844028805761154e-05,
|
| 481 |
+
"loss": 0.7726,
|
| 482 |
+
"step": 7900
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 0.16,
|
| 486 |
+
"learning_rate": 9.842048409681936e-05,
|
| 487 |
+
"loss": 0.7868,
|
| 488 |
+
"step": 8000
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 0.16,
|
| 492 |
+
"learning_rate": 9.840048009601921e-05,
|
| 493 |
+
"loss": 0.7918,
|
| 494 |
+
"step": 8100
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 0.16,
|
| 498 |
+
"learning_rate": 9.838047609521905e-05,
|
| 499 |
+
"loss": 0.7924,
|
| 500 |
+
"step": 8200
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 0.17,
|
| 504 |
+
"learning_rate": 9.836047209441888e-05,
|
| 505 |
+
"loss": 0.8036,
|
| 506 |
+
"step": 8300
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 0.17,
|
| 510 |
+
"learning_rate": 9.834046809361873e-05,
|
| 511 |
+
"loss": 0.7844,
|
| 512 |
+
"step": 8400
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 0.17,
|
| 516 |
+
"learning_rate": 9.832046409281857e-05,
|
| 517 |
+
"loss": 0.8124,
|
| 518 |
+
"step": 8500
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 0.17,
|
| 522 |
+
"learning_rate": 9.830046009201842e-05,
|
| 523 |
+
"loss": 0.837,
|
| 524 |
+
"step": 8600
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 0.17,
|
| 528 |
+
"learning_rate": 9.828045609121826e-05,
|
| 529 |
+
"loss": 0.8315,
|
| 530 |
+
"step": 8700
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 0.18,
|
| 534 |
+
"learning_rate": 9.826045209041809e-05,
|
| 535 |
+
"loss": 0.7738,
|
| 536 |
+
"step": 8800
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 0.18,
|
| 540 |
+
"learning_rate": 9.824044808961794e-05,
|
| 541 |
+
"loss": 0.8164,
|
| 542 |
+
"step": 8900
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 0.18,
|
| 546 |
+
"learning_rate": 9.822044408881777e-05,
|
| 547 |
+
"loss": 0.7529,
|
| 548 |
+
"step": 9000
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 0.18,
|
| 552 |
+
"learning_rate": 9.820044008801761e-05,
|
| 553 |
+
"loss": 0.8037,
|
| 554 |
+
"step": 9100
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 0.18,
|
| 558 |
+
"learning_rate": 9.818043608721746e-05,
|
| 559 |
+
"loss": 0.8537,
|
| 560 |
+
"step": 9200
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 0.19,
|
| 564 |
+
"learning_rate": 9.816043208641729e-05,
|
| 565 |
+
"loss": 0.8241,
|
| 566 |
+
"step": 9300
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 0.19,
|
| 570 |
+
"learning_rate": 9.814042808561713e-05,
|
| 571 |
+
"loss": 0.7729,
|
| 572 |
+
"step": 9400
|
| 573 |
+
},
|
| 574 |
+
{
|
| 575 |
+
"epoch": 0.19,
|
| 576 |
+
"learning_rate": 9.812042408481696e-05,
|
| 577 |
+
"loss": 0.8116,
|
| 578 |
+
"step": 9500
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 0.19,
|
| 582 |
+
"learning_rate": 9.81004200840168e-05,
|
| 583 |
+
"loss": 0.8254,
|
| 584 |
+
"step": 9600
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 0.19,
|
| 588 |
+
"learning_rate": 9.808041608321665e-05,
|
| 589 |
+
"loss": 0.8285,
|
| 590 |
+
"step": 9700
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 0.2,
|
| 594 |
+
"learning_rate": 9.806041208241648e-05,
|
| 595 |
+
"loss": 0.7577,
|
| 596 |
+
"step": 9800
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 0.2,
|
| 600 |
+
"learning_rate": 9.804040808161633e-05,
|
| 601 |
+
"loss": 0.8024,
|
| 602 |
+
"step": 9900
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"epoch": 0.2,
|
| 606 |
+
"learning_rate": 9.802040408081616e-05,
|
| 607 |
+
"loss": 0.8078,
|
| 608 |
+
"step": 10000
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 0.2,
|
| 612 |
+
"eval_loss": 0.740916907787323,
|
| 613 |
+
"eval_runtime": 689.7578,
|
| 614 |
+
"eval_samples_per_second": 7.211,
|
| 615 |
+
"eval_steps_per_second": 0.902,
|
| 616 |
+
"step": 10000
|
| 617 |
+
},
|
| 618 |
+
{
|
| 619 |
+
"epoch": 0.2,
|
| 620 |
+
"learning_rate": 9.8000400080016e-05,
|
| 621 |
+
"loss": 0.7849,
|
| 622 |
+
"step": 10100
|
| 623 |
+
},
|
| 624 |
+
{
|
| 625 |
+
"epoch": 0.2,
|
| 626 |
+
"learning_rate": 9.798039607921585e-05,
|
| 627 |
+
"loss": 0.7931,
|
| 628 |
+
"step": 10200
|
| 629 |
+
},
|
| 630 |
+
{
|
| 631 |
+
"epoch": 0.21,
|
| 632 |
+
"learning_rate": 9.796039207841569e-05,
|
| 633 |
+
"loss": 0.7884,
|
| 634 |
+
"step": 10300
|
| 635 |
+
},
|
| 636 |
+
{
|
| 637 |
+
"epoch": 0.21,
|
| 638 |
+
"learning_rate": 9.794038807761554e-05,
|
| 639 |
+
"loss": 0.7949,
|
| 640 |
+
"step": 10400
|
| 641 |
+
},
|
| 642 |
+
{
|
| 643 |
+
"epoch": 0.21,
|
| 644 |
+
"learning_rate": 9.792038407681537e-05,
|
| 645 |
+
"loss": 0.7916,
|
| 646 |
+
"step": 10500
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"epoch": 0.21,
|
| 650 |
+
"learning_rate": 9.790038007601521e-05,
|
| 651 |
+
"loss": 0.8072,
|
| 652 |
+
"step": 10600
|
| 653 |
+
},
|
| 654 |
+
{
|
| 655 |
+
"epoch": 0.21,
|
| 656 |
+
"learning_rate": 9.788037607521506e-05,
|
| 657 |
+
"loss": 0.7965,
|
| 658 |
+
"step": 10700
|
| 659 |
+
},
|
| 660 |
+
{
|
| 661 |
+
"epoch": 0.22,
|
| 662 |
+
"learning_rate": 9.786037207441489e-05,
|
| 663 |
+
"loss": 0.8039,
|
| 664 |
+
"step": 10800
|
| 665 |
+
},
|
| 666 |
+
{
|
| 667 |
+
"epoch": 0.22,
|
| 668 |
+
"learning_rate": 9.784036807361473e-05,
|
| 669 |
+
"loss": 0.8011,
|
| 670 |
+
"step": 10900
|
| 671 |
+
},
|
| 672 |
+
{
|
| 673 |
+
"epoch": 0.22,
|
| 674 |
+
"learning_rate": 9.782036407281456e-05,
|
| 675 |
+
"loss": 0.8014,
|
| 676 |
+
"step": 11000
|
| 677 |
+
},
|
| 678 |
+
{
|
| 679 |
+
"epoch": 0.22,
|
| 680 |
+
"learning_rate": 9.78003600720144e-05,
|
| 681 |
+
"loss": 0.7835,
|
| 682 |
+
"step": 11100
|
| 683 |
+
},
|
| 684 |
+
{
|
| 685 |
+
"epoch": 0.22,
|
| 686 |
+
"learning_rate": 9.778035607121425e-05,
|
| 687 |
+
"loss": 0.8469,
|
| 688 |
+
"step": 11200
|
| 689 |
+
},
|
| 690 |
+
{
|
| 691 |
+
"epoch": 0.23,
|
| 692 |
+
"learning_rate": 9.776035207041408e-05,
|
| 693 |
+
"loss": 0.8018,
|
| 694 |
+
"step": 11300
|
| 695 |
+
},
|
| 696 |
+
{
|
| 697 |
+
"epoch": 0.23,
|
| 698 |
+
"learning_rate": 9.774034806961393e-05,
|
| 699 |
+
"loss": 0.7839,
|
| 700 |
+
"step": 11400
|
| 701 |
+
},
|
| 702 |
+
{
|
| 703 |
+
"epoch": 0.23,
|
| 704 |
+
"learning_rate": 9.772034406881376e-05,
|
| 705 |
+
"loss": 0.8312,
|
| 706 |
+
"step": 11500
|
| 707 |
+
},
|
| 708 |
+
{
|
| 709 |
+
"epoch": 0.23,
|
| 710 |
+
"learning_rate": 9.770054010802161e-05,
|
| 711 |
+
"loss": 0.7886,
|
| 712 |
+
"step": 11600
|
| 713 |
+
},
|
| 714 |
+
{
|
| 715 |
+
"epoch": 0.23,
|
| 716 |
+
"learning_rate": 9.768053610722146e-05,
|
| 717 |
+
"loss": 0.8096,
|
| 718 |
+
"step": 11700
|
| 719 |
+
},
|
| 720 |
+
{
|
| 721 |
+
"epoch": 0.24,
|
| 722 |
+
"learning_rate": 9.766053210642129e-05,
|
| 723 |
+
"loss": 0.7806,
|
| 724 |
+
"step": 11800
|
| 725 |
+
},
|
| 726 |
+
{
|
| 727 |
+
"epoch": 0.24,
|
| 728 |
+
"learning_rate": 9.764052810562113e-05,
|
| 729 |
+
"loss": 0.8512,
|
| 730 |
+
"step": 11900
|
| 731 |
+
},
|
| 732 |
+
{
|
| 733 |
+
"epoch": 0.24,
|
| 734 |
+
"learning_rate": 9.762052410482096e-05,
|
| 735 |
+
"loss": 0.8475,
|
| 736 |
+
"step": 12000
|
| 737 |
+
},
|
| 738 |
+
{
|
| 739 |
+
"epoch": 0.24,
|
| 740 |
+
"learning_rate": 9.76005201040208e-05,
|
| 741 |
+
"loss": 0.749,
|
| 742 |
+
"step": 12100
|
| 743 |
+
},
|
| 744 |
+
{
|
| 745 |
+
"epoch": 0.24,
|
| 746 |
+
"learning_rate": 9.758051610322065e-05,
|
| 747 |
+
"loss": 0.8245,
|
| 748 |
+
"step": 12200
|
| 749 |
+
},
|
| 750 |
+
{
|
| 751 |
+
"epoch": 0.25,
|
| 752 |
+
"learning_rate": 9.756051210242048e-05,
|
| 753 |
+
"loss": 0.7966,
|
| 754 |
+
"step": 12300
|
| 755 |
+
},
|
| 756 |
+
{
|
| 757 |
+
"epoch": 0.25,
|
| 758 |
+
"learning_rate": 9.754050810162033e-05,
|
| 759 |
+
"loss": 0.773,
|
| 760 |
+
"step": 12400
|
| 761 |
+
},
|
| 762 |
+
{
|
| 763 |
+
"epoch": 0.25,
|
| 764 |
+
"learning_rate": 9.752050410082017e-05,
|
| 765 |
+
"loss": 0.7876,
|
| 766 |
+
"step": 12500
|
| 767 |
+
},
|
| 768 |
+
{
|
| 769 |
+
"epoch": 0.25,
|
| 770 |
+
"learning_rate": 9.750050010002001e-05,
|
| 771 |
+
"loss": 0.7982,
|
| 772 |
+
"step": 12600
|
| 773 |
+
},
|
| 774 |
+
{
|
| 775 |
+
"epoch": 0.25,
|
| 776 |
+
"learning_rate": 9.748049609921986e-05,
|
| 777 |
+
"loss": 0.7509,
|
| 778 |
+
"step": 12700
|
| 779 |
+
},
|
| 780 |
+
{
|
| 781 |
+
"epoch": 0.26,
|
| 782 |
+
"learning_rate": 9.746049209841969e-05,
|
| 783 |
+
"loss": 0.7815,
|
| 784 |
+
"step": 12800
|
| 785 |
+
},
|
| 786 |
+
{
|
| 787 |
+
"epoch": 0.26,
|
| 788 |
+
"learning_rate": 9.744048809761953e-05,
|
| 789 |
+
"loss": 0.8516,
|
| 790 |
+
"step": 12900
|
| 791 |
+
},
|
| 792 |
+
{
|
| 793 |
+
"epoch": 0.26,
|
| 794 |
+
"learning_rate": 9.742048409681937e-05,
|
| 795 |
+
"loss": 0.7684,
|
| 796 |
+
"step": 13000
|
| 797 |
+
},
|
| 798 |
+
{
|
| 799 |
+
"epoch": 0.26,
|
| 800 |
+
"learning_rate": 9.740048009601921e-05,
|
| 801 |
+
"loss": 0.7889,
|
| 802 |
+
"step": 13100
|
| 803 |
+
},
|
| 804 |
+
{
|
| 805 |
+
"epoch": 0.26,
|
| 806 |
+
"learning_rate": 9.738047609521905e-05,
|
| 807 |
+
"loss": 0.7882,
|
| 808 |
+
"step": 13200
|
| 809 |
+
},
|
| 810 |
+
{
|
| 811 |
+
"epoch": 0.27,
|
| 812 |
+
"learning_rate": 9.736047209441889e-05,
|
| 813 |
+
"loss": 0.8196,
|
| 814 |
+
"step": 13300
|
| 815 |
+
},
|
| 816 |
+
{
|
| 817 |
+
"epoch": 0.27,
|
| 818 |
+
"learning_rate": 9.734046809361873e-05,
|
| 819 |
+
"loss": 0.7889,
|
| 820 |
+
"step": 13400
|
| 821 |
+
},
|
| 822 |
+
{
|
| 823 |
+
"epoch": 0.27,
|
| 824 |
+
"learning_rate": 9.732046409281856e-05,
|
| 825 |
+
"loss": 0.7931,
|
| 826 |
+
"step": 13500
|
| 827 |
+
},
|
| 828 |
+
{
|
| 829 |
+
"epoch": 0.27,
|
| 830 |
+
"learning_rate": 9.73004600920184e-05,
|
| 831 |
+
"loss": 0.8172,
|
| 832 |
+
"step": 13600
|
| 833 |
+
},
|
| 834 |
+
{
|
| 835 |
+
"epoch": 0.27,
|
| 836 |
+
"learning_rate": 9.728045609121825e-05,
|
| 837 |
+
"loss": 0.8364,
|
| 838 |
+
"step": 13700
|
| 839 |
+
},
|
| 840 |
+
{
|
| 841 |
+
"epoch": 0.28,
|
| 842 |
+
"learning_rate": 9.726045209041808e-05,
|
| 843 |
+
"loss": 0.822,
|
| 844 |
+
"step": 13800
|
| 845 |
+
},
|
| 846 |
+
{
|
| 847 |
+
"epoch": 0.28,
|
| 848 |
+
"learning_rate": 9.724044808961793e-05,
|
| 849 |
+
"loss": 0.8546,
|
| 850 |
+
"step": 13900
|
| 851 |
+
},
|
| 852 |
+
{
|
| 853 |
+
"epoch": 0.28,
|
| 854 |
+
"learning_rate": 9.722044408881777e-05,
|
| 855 |
+
"loss": 0.7977,
|
| 856 |
+
"step": 14000
|
| 857 |
+
},
|
| 858 |
+
{
|
| 859 |
+
"epoch": 0.28,
|
| 860 |
+
"learning_rate": 9.720044008801761e-05,
|
| 861 |
+
"loss": 0.7591,
|
| 862 |
+
"step": 14100
|
| 863 |
+
},
|
| 864 |
+
{
|
| 865 |
+
"epoch": 0.28,
|
| 866 |
+
"learning_rate": 9.718063612722545e-05,
|
| 867 |
+
"loss": 0.7787,
|
| 868 |
+
"step": 14200
|
| 869 |
+
},
|
| 870 |
+
{
|
| 871 |
+
"epoch": 0.29,
|
| 872 |
+
"learning_rate": 9.716063212642529e-05,
|
| 873 |
+
"loss": 0.8361,
|
| 874 |
+
"step": 14300
|
| 875 |
+
},
|
| 876 |
+
{
|
| 877 |
+
"epoch": 0.29,
|
| 878 |
+
"learning_rate": 9.714062812562513e-05,
|
| 879 |
+
"loss": 0.8153,
|
| 880 |
+
"step": 14400
|
| 881 |
+
},
|
| 882 |
+
{
|
| 883 |
+
"epoch": 0.29,
|
| 884 |
+
"learning_rate": 9.712062412482496e-05,
|
| 885 |
+
"loss": 0.776,
|
| 886 |
+
"step": 14500
|
| 887 |
+
},
|
| 888 |
+
{
|
| 889 |
+
"epoch": 0.29,
|
| 890 |
+
"learning_rate": 9.71006201240248e-05,
|
| 891 |
+
"loss": 0.7491,
|
| 892 |
+
"step": 14600
|
| 893 |
+
},
|
| 894 |
+
{
|
| 895 |
+
"epoch": 0.29,
|
| 896 |
+
"learning_rate": 9.708061612322465e-05,
|
| 897 |
+
"loss": 0.7975,
|
| 898 |
+
"step": 14700
|
| 899 |
+
},
|
| 900 |
+
{
|
| 901 |
+
"epoch": 0.3,
|
| 902 |
+
"learning_rate": 9.70606121224245e-05,
|
| 903 |
+
"loss": 0.7637,
|
| 904 |
+
"step": 14800
|
| 905 |
+
},
|
| 906 |
+
{
|
| 907 |
+
"epoch": 0.3,
|
| 908 |
+
"learning_rate": 9.704060812162434e-05,
|
| 909 |
+
"loss": 0.7977,
|
| 910 |
+
"step": 14900
|
| 911 |
+
},
|
| 912 |
+
{
|
| 913 |
+
"epoch": 0.3,
|
| 914 |
+
"learning_rate": 9.702060412082417e-05,
|
| 915 |
+
"loss": 0.8041,
|
| 916 |
+
"step": 15000
|
| 917 |
+
},
|
| 918 |
+
{
|
| 919 |
+
"epoch": 0.3,
|
| 920 |
+
"learning_rate": 9.700060012002401e-05,
|
| 921 |
+
"loss": 0.7997,
|
| 922 |
+
"step": 15100
|
| 923 |
+
},
|
| 924 |
+
{
|
| 925 |
+
"epoch": 0.3,
|
| 926 |
+
"learning_rate": 9.698059611922386e-05,
|
| 927 |
+
"loss": 0.8221,
|
| 928 |
+
"step": 15200
|
| 929 |
+
},
|
| 930 |
+
{
|
| 931 |
+
"epoch": 0.31,
|
| 932 |
+
"learning_rate": 9.696059211842369e-05,
|
| 933 |
+
"loss": 0.7657,
|
| 934 |
+
"step": 15300
|
| 935 |
+
},
|
| 936 |
+
{
|
| 937 |
+
"epoch": 0.31,
|
| 938 |
+
"learning_rate": 9.694058811762353e-05,
|
| 939 |
+
"loss": 0.8091,
|
| 940 |
+
"step": 15400
|
| 941 |
+
},
|
| 942 |
+
{
|
| 943 |
+
"epoch": 0.31,
|
| 944 |
+
"learning_rate": 9.692058411682336e-05,
|
| 945 |
+
"loss": 0.7692,
|
| 946 |
+
"step": 15500
|
| 947 |
+
},
|
| 948 |
+
{
|
| 949 |
+
"epoch": 0.31,
|
| 950 |
+
"learning_rate": 9.690058011602321e-05,
|
| 951 |
+
"loss": 0.7988,
|
| 952 |
+
"step": 15600
|
| 953 |
+
},
|
| 954 |
+
{
|
| 955 |
+
"epoch": 0.31,
|
| 956 |
+
"learning_rate": 9.688057611522305e-05,
|
| 957 |
+
"loss": 0.7457,
|
| 958 |
+
"step": 15700
|
| 959 |
+
},
|
| 960 |
+
{
|
| 961 |
+
"epoch": 0.32,
|
| 962 |
+
"learning_rate": 9.686057211442288e-05,
|
| 963 |
+
"loss": 0.8223,
|
| 964 |
+
"step": 15800
|
| 965 |
+
},
|
| 966 |
+
{
|
| 967 |
+
"epoch": 0.32,
|
| 968 |
+
"learning_rate": 9.684056811362273e-05,
|
| 969 |
+
"loss": 0.8366,
|
| 970 |
+
"step": 15900
|
| 971 |
+
},
|
| 972 |
+
{
|
| 973 |
+
"epoch": 0.32,
|
| 974 |
+
"learning_rate": 9.682056411282256e-05,
|
| 975 |
+
"loss": 0.7959,
|
| 976 |
+
"step": 16000
|
| 977 |
+
},
|
| 978 |
+
{
|
| 979 |
+
"epoch": 0.32,
|
| 980 |
+
"learning_rate": 9.680076015203041e-05,
|
| 981 |
+
"loss": 0.7944,
|
| 982 |
+
"step": 16100
|
| 983 |
+
},
|
| 984 |
+
{
|
| 985 |
+
"epoch": 0.32,
|
| 986 |
+
"learning_rate": 9.678075615123024e-05,
|
| 987 |
+
"loss": 0.7645,
|
| 988 |
+
"step": 16200
|
| 989 |
+
},
|
| 990 |
+
{
|
| 991 |
+
"epoch": 0.33,
|
| 992 |
+
"learning_rate": 9.676075215043009e-05,
|
| 993 |
+
"loss": 0.8111,
|
| 994 |
+
"step": 16300
|
| 995 |
+
},
|
| 996 |
+
{
|
| 997 |
+
"epoch": 0.33,
|
| 998 |
+
"learning_rate": 9.674074814962993e-05,
|
| 999 |
+
"loss": 0.7907,
|
| 1000 |
+
"step": 16400
|
| 1001 |
+
},
|
| 1002 |
+
{
|
| 1003 |
+
"epoch": 0.33,
|
| 1004 |
+
"learning_rate": 9.672074414882976e-05,
|
| 1005 |
+
"loss": 0.8108,
|
| 1006 |
+
"step": 16500
|
| 1007 |
+
},
|
| 1008 |
+
{
|
| 1009 |
+
"epoch": 0.33,
|
| 1010 |
+
"learning_rate": 9.670074014802961e-05,
|
| 1011 |
+
"loss": 0.8119,
|
| 1012 |
+
"step": 16600
|
| 1013 |
+
},
|
| 1014 |
+
{
|
| 1015 |
+
"epoch": 0.33,
|
| 1016 |
+
"learning_rate": 9.668073614722945e-05,
|
| 1017 |
+
"loss": 0.8109,
|
| 1018 |
+
"step": 16700
|
| 1019 |
+
},
|
| 1020 |
+
{
|
| 1021 |
+
"epoch": 0.34,
|
| 1022 |
+
"learning_rate": 9.666073214642928e-05,
|
| 1023 |
+
"loss": 0.7862,
|
| 1024 |
+
"step": 16800
|
| 1025 |
+
},
|
| 1026 |
+
{
|
| 1027 |
+
"epoch": 0.34,
|
| 1028 |
+
"learning_rate": 9.664072814562913e-05,
|
| 1029 |
+
"loss": 0.7886,
|
| 1030 |
+
"step": 16900
|
| 1031 |
+
},
|
| 1032 |
+
{
|
| 1033 |
+
"epoch": 0.34,
|
| 1034 |
+
"learning_rate": 9.662072414482897e-05,
|
| 1035 |
+
"loss": 0.786,
|
| 1036 |
+
"step": 17000
|
| 1037 |
+
},
|
| 1038 |
+
{
|
| 1039 |
+
"epoch": 0.34,
|
| 1040 |
+
"learning_rate": 9.660072014402882e-05,
|
| 1041 |
+
"loss": 0.8014,
|
| 1042 |
+
"step": 17100
|
| 1043 |
+
},
|
| 1044 |
+
{
|
| 1045 |
+
"epoch": 0.34,
|
| 1046 |
+
"learning_rate": 9.658071614322865e-05,
|
| 1047 |
+
"loss": 0.7486,
|
| 1048 |
+
"step": 17200
|
| 1049 |
+
},
|
| 1050 |
+
{
|
| 1051 |
+
"epoch": 0.35,
|
| 1052 |
+
"learning_rate": 9.65607121424285e-05,
|
| 1053 |
+
"loss": 0.7688,
|
| 1054 |
+
"step": 17300
|
| 1055 |
+
},
|
| 1056 |
+
{
|
| 1057 |
+
"epoch": 0.35,
|
| 1058 |
+
"learning_rate": 9.654070814162834e-05,
|
| 1059 |
+
"loss": 0.7815,
|
| 1060 |
+
"step": 17400
|
| 1061 |
+
},
|
| 1062 |
+
{
|
| 1063 |
+
"epoch": 0.35,
|
| 1064 |
+
"learning_rate": 9.652070414082817e-05,
|
| 1065 |
+
"loss": 0.7857,
|
| 1066 |
+
"step": 17500
|
| 1067 |
+
},
|
| 1068 |
+
{
|
| 1069 |
+
"epoch": 0.35,
|
| 1070 |
+
"learning_rate": 9.650070014002801e-05,
|
| 1071 |
+
"loss": 0.777,
|
| 1072 |
+
"step": 17600
|
| 1073 |
+
},
|
| 1074 |
+
{
|
| 1075 |
+
"epoch": 0.35,
|
| 1076 |
+
"learning_rate": 9.648069613922786e-05,
|
| 1077 |
+
"loss": 0.8186,
|
| 1078 |
+
"step": 17700
|
| 1079 |
+
},
|
| 1080 |
+
{
|
| 1081 |
+
"epoch": 0.36,
|
| 1082 |
+
"learning_rate": 9.646069213842769e-05,
|
| 1083 |
+
"loss": 0.7765,
|
| 1084 |
+
"step": 17800
|
| 1085 |
+
},
|
| 1086 |
+
{
|
| 1087 |
+
"epoch": 0.36,
|
| 1088 |
+
"learning_rate": 9.644068813762753e-05,
|
| 1089 |
+
"loss": 0.8268,
|
| 1090 |
+
"step": 17900
|
| 1091 |
+
},
|
| 1092 |
+
{
|
| 1093 |
+
"epoch": 0.36,
|
| 1094 |
+
"learning_rate": 9.642068413682736e-05,
|
| 1095 |
+
"loss": 0.7864,
|
| 1096 |
+
"step": 18000
|
| 1097 |
+
},
|
| 1098 |
+
{
|
| 1099 |
+
"epoch": 0.36,
|
| 1100 |
+
"learning_rate": 9.640068013602721e-05,
|
| 1101 |
+
"loss": 0.7741,
|
| 1102 |
+
"step": 18100
|
| 1103 |
+
},
|
| 1104 |
+
{
|
| 1105 |
+
"epoch": 0.36,
|
| 1106 |
+
"learning_rate": 9.638067613522705e-05,
|
| 1107 |
+
"loss": 0.831,
|
| 1108 |
+
"step": 18200
|
| 1109 |
+
},
|
| 1110 |
+
{
|
| 1111 |
+
"epoch": 0.37,
|
| 1112 |
+
"learning_rate": 9.636067213442688e-05,
|
| 1113 |
+
"loss": 0.7912,
|
| 1114 |
+
"step": 18300
|
| 1115 |
+
},
|
| 1116 |
+
{
|
| 1117 |
+
"epoch": 0.37,
|
| 1118 |
+
"learning_rate": 9.634066813362673e-05,
|
| 1119 |
+
"loss": 0.7864,
|
| 1120 |
+
"step": 18400
|
| 1121 |
+
},
|
| 1122 |
+
{
|
| 1123 |
+
"epoch": 0.37,
|
| 1124 |
+
"learning_rate": 9.632066413282656e-05,
|
| 1125 |
+
"loss": 0.7639,
|
| 1126 |
+
"step": 18500
|
| 1127 |
+
},
|
| 1128 |
+
{
|
| 1129 |
+
"epoch": 0.37,
|
| 1130 |
+
"learning_rate": 9.63006601320264e-05,
|
| 1131 |
+
"loss": 0.7874,
|
| 1132 |
+
"step": 18600
|
| 1133 |
+
},
|
| 1134 |
+
{
|
| 1135 |
+
"epoch": 0.37,
|
| 1136 |
+
"learning_rate": 9.628065613122625e-05,
|
| 1137 |
+
"loss": 0.7642,
|
| 1138 |
+
"step": 18700
|
| 1139 |
+
},
|
| 1140 |
+
{
|
| 1141 |
+
"epoch": 0.38,
|
| 1142 |
+
"learning_rate": 9.626065213042609e-05,
|
| 1143 |
+
"loss": 0.756,
|
| 1144 |
+
"step": 18800
|
| 1145 |
+
},
|
| 1146 |
+
{
|
| 1147 |
+
"epoch": 0.38,
|
| 1148 |
+
"learning_rate": 9.624064812962594e-05,
|
| 1149 |
+
"loss": 0.8036,
|
| 1150 |
+
"step": 18900
|
| 1151 |
+
},
|
| 1152 |
+
{
|
| 1153 |
+
"epoch": 0.38,
|
| 1154 |
+
"learning_rate": 9.622064412882577e-05,
|
| 1155 |
+
"loss": 0.7537,
|
| 1156 |
+
"step": 19000
|
| 1157 |
+
},
|
| 1158 |
+
{
|
| 1159 |
+
"epoch": 0.38,
|
| 1160 |
+
"learning_rate": 9.620064012802561e-05,
|
| 1161 |
+
"loss": 0.8193,
|
| 1162 |
+
"step": 19100
|
| 1163 |
+
},
|
| 1164 |
+
{
|
| 1165 |
+
"epoch": 0.38,
|
| 1166 |
+
"learning_rate": 9.618063612722546e-05,
|
| 1167 |
+
"loss": 0.8036,
|
| 1168 |
+
"step": 19200
|
| 1169 |
+
},
|
| 1170 |
+
{
|
| 1171 |
+
"epoch": 0.39,
|
| 1172 |
+
"learning_rate": 9.616063212642529e-05,
|
| 1173 |
+
"loss": 0.784,
|
| 1174 |
+
"step": 19300
|
| 1175 |
+
},
|
| 1176 |
+
{
|
| 1177 |
+
"epoch": 0.39,
|
| 1178 |
+
"learning_rate": 9.614062812562513e-05,
|
| 1179 |
+
"loss": 0.7861,
|
| 1180 |
+
"step": 19400
|
| 1181 |
+
},
|
| 1182 |
+
{
|
| 1183 |
+
"epoch": 0.39,
|
| 1184 |
+
"learning_rate": 9.612062412482496e-05,
|
| 1185 |
+
"loss": 0.7882,
|
| 1186 |
+
"step": 19500
|
| 1187 |
+
},
|
| 1188 |
+
{
|
| 1189 |
+
"epoch": 0.39,
|
| 1190 |
+
"learning_rate": 9.610062012402481e-05,
|
| 1191 |
+
"loss": 0.8048,
|
| 1192 |
+
"step": 19600
|
| 1193 |
+
},
|
| 1194 |
+
{
|
| 1195 |
+
"epoch": 0.39,
|
| 1196 |
+
"learning_rate": 9.608061612322465e-05,
|
| 1197 |
+
"loss": 0.8544,
|
| 1198 |
+
"step": 19700
|
| 1199 |
+
},
|
| 1200 |
+
{
|
| 1201 |
+
"epoch": 0.4,
|
| 1202 |
+
"learning_rate": 9.606061212242448e-05,
|
| 1203 |
+
"loss": 0.7827,
|
| 1204 |
+
"step": 19800
|
| 1205 |
+
},
|
| 1206 |
+
{
|
| 1207 |
+
"epoch": 0.4,
|
| 1208 |
+
"learning_rate": 9.604060812162433e-05,
|
| 1209 |
+
"loss": 0.7641,
|
| 1210 |
+
"step": 19900
|
| 1211 |
+
},
|
| 1212 |
+
{
|
| 1213 |
+
"epoch": 0.4,
|
| 1214 |
+
"learning_rate": 9.602060412082416e-05,
|
| 1215 |
+
"loss": 0.7146,
|
| 1216 |
+
"step": 20000
|
| 1217 |
+
},
|
| 1218 |
+
{
|
| 1219 |
+
"epoch": 0.4,
|
| 1220 |
+
"eval_loss": 0.7286937832832336,
|
| 1221 |
+
"eval_runtime": 689.1627,
|
| 1222 |
+
"eval_samples_per_second": 7.217,
|
| 1223 |
+
"eval_steps_per_second": 0.903,
|
| 1224 |
+
"step": 20000
|
| 1225 |
+
},
|
| 1226 |
+
{
|
| 1227 |
+
"epoch": 0.4,
|
| 1228 |
+
"learning_rate": 9.6000600120024e-05,
|
| 1229 |
+
"loss": 0.7801,
|
| 1230 |
+
"step": 20100
|
| 1231 |
+
},
|
| 1232 |
+
{
|
| 1233 |
+
"epoch": 0.4,
|
| 1234 |
+
"learning_rate": 9.598059611922385e-05,
|
| 1235 |
+
"loss": 0.7803,
|
| 1236 |
+
"step": 20200
|
| 1237 |
+
},
|
| 1238 |
+
{
|
| 1239 |
+
"epoch": 0.41,
|
| 1240 |
+
"learning_rate": 9.596059211842369e-05,
|
| 1241 |
+
"loss": 0.7981,
|
| 1242 |
+
"step": 20300
|
| 1243 |
+
},
|
| 1244 |
+
{
|
| 1245 |
+
"epoch": 0.41,
|
| 1246 |
+
"learning_rate": 9.594058811762354e-05,
|
| 1247 |
+
"loss": 0.755,
|
| 1248 |
+
"step": 20400
|
| 1249 |
+
},
|
| 1250 |
+
{
|
| 1251 |
+
"epoch": 0.41,
|
| 1252 |
+
"learning_rate": 9.592058411682337e-05,
|
| 1253 |
+
"loss": 0.7743,
|
| 1254 |
+
"step": 20500
|
| 1255 |
+
},
|
| 1256 |
+
{
|
| 1257 |
+
"epoch": 0.41,
|
| 1258 |
+
"learning_rate": 9.590098019603922e-05,
|
| 1259 |
+
"loss": 0.7358,
|
| 1260 |
+
"step": 20600
|
| 1261 |
+
},
|
| 1262 |
+
{
|
| 1263 |
+
"epoch": 0.41,
|
| 1264 |
+
"learning_rate": 9.588097619523905e-05,
|
| 1265 |
+
"loss": 0.7851,
|
| 1266 |
+
"step": 20700
|
| 1267 |
+
},
|
| 1268 |
+
{
|
| 1269 |
+
"epoch": 0.42,
|
| 1270 |
+
"learning_rate": 9.586097219443889e-05,
|
| 1271 |
+
"loss": 0.7746,
|
| 1272 |
+
"step": 20800
|
| 1273 |
+
},
|
| 1274 |
+
{
|
| 1275 |
+
"epoch": 0.42,
|
| 1276 |
+
"learning_rate": 9.584096819363874e-05,
|
| 1277 |
+
"loss": 0.8246,
|
| 1278 |
+
"step": 20900
|
| 1279 |
+
},
|
| 1280 |
+
{
|
| 1281 |
+
"epoch": 0.42,
|
| 1282 |
+
"learning_rate": 9.582096419283857e-05,
|
| 1283 |
+
"loss": 0.7833,
|
| 1284 |
+
"step": 21000
|
| 1285 |
+
},
|
| 1286 |
+
{
|
| 1287 |
+
"epoch": 0.42,
|
| 1288 |
+
"learning_rate": 9.580096019203841e-05,
|
| 1289 |
+
"loss": 0.7666,
|
| 1290 |
+
"step": 21100
|
| 1291 |
+
},
|
| 1292 |
+
{
|
| 1293 |
+
"epoch": 0.42,
|
| 1294 |
+
"learning_rate": 9.578095619123824e-05,
|
| 1295 |
+
"loss": 0.8575,
|
| 1296 |
+
"step": 21200
|
| 1297 |
+
},
|
| 1298 |
+
{
|
| 1299 |
+
"epoch": 0.43,
|
| 1300 |
+
"learning_rate": 9.576095219043809e-05,
|
| 1301 |
+
"loss": 0.7934,
|
| 1302 |
+
"step": 21300
|
| 1303 |
+
},
|
| 1304 |
+
{
|
| 1305 |
+
"epoch": 0.43,
|
| 1306 |
+
"learning_rate": 9.574094818963793e-05,
|
| 1307 |
+
"loss": 0.7473,
|
| 1308 |
+
"step": 21400
|
| 1309 |
+
},
|
| 1310 |
+
{
|
| 1311 |
+
"epoch": 0.43,
|
| 1312 |
+
"learning_rate": 9.572094418883776e-05,
|
| 1313 |
+
"loss": 0.7455,
|
| 1314 |
+
"step": 21500
|
| 1315 |
+
},
|
| 1316 |
+
{
|
| 1317 |
+
"epoch": 0.43,
|
| 1318 |
+
"learning_rate": 9.570094018803761e-05,
|
| 1319 |
+
"loss": 0.7527,
|
| 1320 |
+
"step": 21600
|
| 1321 |
+
},
|
| 1322 |
+
{
|
| 1323 |
+
"epoch": 0.43,
|
| 1324 |
+
"learning_rate": 9.568093618723745e-05,
|
| 1325 |
+
"loss": 0.7906,
|
| 1326 |
+
"step": 21700
|
| 1327 |
+
},
|
| 1328 |
+
{
|
| 1329 |
+
"epoch": 0.44,
|
| 1330 |
+
"learning_rate": 9.56609321864373e-05,
|
| 1331 |
+
"loss": 0.8384,
|
| 1332 |
+
"step": 21800
|
| 1333 |
+
},
|
| 1334 |
+
{
|
| 1335 |
+
"epoch": 0.44,
|
| 1336 |
+
"learning_rate": 9.564092818563714e-05,
|
| 1337 |
+
"loss": 0.7897,
|
| 1338 |
+
"step": 21900
|
| 1339 |
+
},
|
| 1340 |
+
{
|
| 1341 |
+
"epoch": 0.44,
|
| 1342 |
+
"learning_rate": 9.562092418483697e-05,
|
| 1343 |
+
"loss": 0.7459,
|
| 1344 |
+
"step": 22000
|
| 1345 |
+
},
|
| 1346 |
+
{
|
| 1347 |
+
"epoch": 0.44,
|
| 1348 |
+
"learning_rate": 9.560092018403682e-05,
|
| 1349 |
+
"loss": 0.7838,
|
| 1350 |
+
"step": 22100
|
| 1351 |
+
},
|
| 1352 |
+
{
|
| 1353 |
+
"epoch": 0.44,
|
| 1354 |
+
"learning_rate": 9.558091618323666e-05,
|
| 1355 |
+
"loss": 0.7745,
|
| 1356 |
+
"step": 22200
|
| 1357 |
+
},
|
| 1358 |
+
{
|
| 1359 |
+
"epoch": 0.45,
|
| 1360 |
+
"learning_rate": 9.556091218243649e-05,
|
| 1361 |
+
"loss": 0.7985,
|
| 1362 |
+
"step": 22300
|
| 1363 |
+
},
|
| 1364 |
+
{
|
| 1365 |
+
"epoch": 0.45,
|
| 1366 |
+
"learning_rate": 9.554090818163634e-05,
|
| 1367 |
+
"loss": 0.7945,
|
| 1368 |
+
"step": 22400
|
| 1369 |
+
},
|
| 1370 |
+
{
|
| 1371 |
+
"epoch": 0.45,
|
| 1372 |
+
"learning_rate": 9.552090418083617e-05,
|
| 1373 |
+
"loss": 0.7985,
|
| 1374 |
+
"step": 22500
|
| 1375 |
+
},
|
| 1376 |
+
{
|
| 1377 |
+
"epoch": 0.45,
|
| 1378 |
+
"learning_rate": 9.550090018003601e-05,
|
| 1379 |
+
"loss": 0.7785,
|
| 1380 |
+
"step": 22600
|
| 1381 |
+
},
|
| 1382 |
+
{
|
| 1383 |
+
"epoch": 0.45,
|
| 1384 |
+
"learning_rate": 9.548089617923586e-05,
|
| 1385 |
+
"loss": 0.7642,
|
| 1386 |
+
"step": 22700
|
| 1387 |
+
},
|
| 1388 |
+
{
|
| 1389 |
+
"epoch": 0.46,
|
| 1390 |
+
"learning_rate": 9.546089217843569e-05,
|
| 1391 |
+
"loss": 0.7953,
|
| 1392 |
+
"step": 22800
|
| 1393 |
+
},
|
| 1394 |
+
{
|
| 1395 |
+
"epoch": 0.46,
|
| 1396 |
+
"learning_rate": 9.544088817763553e-05,
|
| 1397 |
+
"loss": 0.7848,
|
| 1398 |
+
"step": 22900
|
| 1399 |
+
},
|
| 1400 |
+
{
|
| 1401 |
+
"epoch": 0.46,
|
| 1402 |
+
"learning_rate": 9.542088417683536e-05,
|
| 1403 |
+
"loss": 0.7388,
|
| 1404 |
+
"step": 23000
|
| 1405 |
+
},
|
| 1406 |
+
{
|
| 1407 |
+
"epoch": 0.46,
|
| 1408 |
+
"learning_rate": 9.540088017603521e-05,
|
| 1409 |
+
"loss": 0.7947,
|
| 1410 |
+
"step": 23100
|
| 1411 |
+
},
|
| 1412 |
+
{
|
| 1413 |
+
"epoch": 0.46,
|
| 1414 |
+
"learning_rate": 9.538087617523505e-05,
|
| 1415 |
+
"loss": 0.7464,
|
| 1416 |
+
"step": 23200
|
| 1417 |
+
},
|
| 1418 |
+
{
|
| 1419 |
+
"epoch": 0.47,
|
| 1420 |
+
"learning_rate": 9.53608721744349e-05,
|
| 1421 |
+
"loss": 0.7696,
|
| 1422 |
+
"step": 23300
|
| 1423 |
+
},
|
| 1424 |
+
{
|
| 1425 |
+
"epoch": 0.47,
|
| 1426 |
+
"learning_rate": 9.534086817363474e-05,
|
| 1427 |
+
"loss": 0.7563,
|
| 1428 |
+
"step": 23400
|
| 1429 |
+
},
|
| 1430 |
+
{
|
| 1431 |
+
"epoch": 0.47,
|
| 1432 |
+
"learning_rate": 9.532086417283457e-05,
|
| 1433 |
+
"loss": 0.7522,
|
| 1434 |
+
"step": 23500
|
| 1435 |
+
},
|
| 1436 |
+
{
|
| 1437 |
+
"epoch": 0.47,
|
| 1438 |
+
"learning_rate": 9.530086017203442e-05,
|
| 1439 |
+
"loss": 0.772,
|
| 1440 |
+
"step": 23600
|
| 1441 |
+
},
|
| 1442 |
+
{
|
| 1443 |
+
"epoch": 0.47,
|
| 1444 |
+
"learning_rate": 9.528085617123426e-05,
|
| 1445 |
+
"loss": 0.7654,
|
| 1446 |
+
"step": 23700
|
| 1447 |
+
},
|
| 1448 |
+
{
|
| 1449 |
+
"epoch": 0.48,
|
| 1450 |
+
"learning_rate": 9.526085217043409e-05,
|
| 1451 |
+
"loss": 0.759,
|
| 1452 |
+
"step": 23800
|
| 1453 |
+
},
|
| 1454 |
+
{
|
| 1455 |
+
"epoch": 0.48,
|
| 1456 |
+
"learning_rate": 9.524084816963394e-05,
|
| 1457 |
+
"loss": 0.8032,
|
| 1458 |
+
"step": 23900
|
| 1459 |
+
},
|
| 1460 |
+
{
|
| 1461 |
+
"epoch": 0.48,
|
| 1462 |
+
"learning_rate": 9.522084416883377e-05,
|
| 1463 |
+
"loss": 0.7638,
|
| 1464 |
+
"step": 24000
|
| 1465 |
+
},
|
| 1466 |
+
{
|
| 1467 |
+
"epoch": 0.48,
|
| 1468 |
+
"learning_rate": 9.520084016803361e-05,
|
| 1469 |
+
"loss": 0.7974,
|
| 1470 |
+
"step": 24100
|
| 1471 |
+
},
|
| 1472 |
+
{
|
| 1473 |
+
"epoch": 0.48,
|
| 1474 |
+
"learning_rate": 9.518083616723346e-05,
|
| 1475 |
+
"loss": 0.7647,
|
| 1476 |
+
"step": 24200
|
| 1477 |
+
},
|
| 1478 |
+
{
|
| 1479 |
+
"epoch": 0.49,
|
| 1480 |
+
"learning_rate": 9.516083216643329e-05,
|
| 1481 |
+
"loss": 0.8016,
|
| 1482 |
+
"step": 24300
|
| 1483 |
+
},
|
| 1484 |
+
{
|
| 1485 |
+
"epoch": 0.49,
|
| 1486 |
+
"learning_rate": 9.514082816563313e-05,
|
| 1487 |
+
"loss": 0.7749,
|
| 1488 |
+
"step": 24400
|
| 1489 |
+
},
|
| 1490 |
+
{
|
| 1491 |
+
"epoch": 0.49,
|
| 1492 |
+
"learning_rate": 9.512082416483296e-05,
|
| 1493 |
+
"loss": 0.7573,
|
| 1494 |
+
"step": 24500
|
| 1495 |
+
},
|
| 1496 |
+
{
|
| 1497 |
+
"epoch": 0.49,
|
| 1498 |
+
"learning_rate": 9.51008201640328e-05,
|
| 1499 |
+
"loss": 0.7476,
|
| 1500 |
+
"step": 24600
|
| 1501 |
+
},
|
| 1502 |
+
{
|
| 1503 |
+
"epoch": 0.49,
|
| 1504 |
+
"learning_rate": 9.508081616323265e-05,
|
| 1505 |
+
"loss": 0.7848,
|
| 1506 |
+
"step": 24700
|
| 1507 |
+
},
|
| 1508 |
+
{
|
| 1509 |
+
"epoch": 0.5,
|
| 1510 |
+
"learning_rate": 9.50608121624325e-05,
|
| 1511 |
+
"loss": 0.7994,
|
| 1512 |
+
"step": 24800
|
| 1513 |
+
},
|
| 1514 |
+
{
|
| 1515 |
+
"epoch": 0.5,
|
| 1516 |
+
"learning_rate": 9.504100820164034e-05,
|
| 1517 |
+
"loss": 0.8056,
|
| 1518 |
+
"step": 24900
|
| 1519 |
+
},
|
| 1520 |
+
{
|
| 1521 |
+
"epoch": 0.5,
|
| 1522 |
+
"learning_rate": 9.502100420084017e-05,
|
| 1523 |
+
"loss": 0.8066,
|
| 1524 |
+
"step": 25000
|
| 1525 |
+
},
|
| 1526 |
+
{
|
| 1527 |
+
"epoch": 0.5,
|
| 1528 |
+
"learning_rate": 9.500100020004001e-05,
|
| 1529 |
+
"loss": 0.7506,
|
| 1530 |
+
"step": 25100
|
| 1531 |
+
},
|
| 1532 |
+
{
|
| 1533 |
+
"epoch": 0.5,
|
| 1534 |
+
"learning_rate": 9.498099619923986e-05,
|
| 1535 |
+
"loss": 0.8018,
|
| 1536 |
+
"step": 25200
|
| 1537 |
+
},
|
| 1538 |
+
{
|
| 1539 |
+
"epoch": 0.51,
|
| 1540 |
+
"learning_rate": 9.496099219843969e-05,
|
| 1541 |
+
"loss": 0.7451,
|
| 1542 |
+
"step": 25300
|
| 1543 |
+
},
|
| 1544 |
+
{
|
| 1545 |
+
"epoch": 0.51,
|
| 1546 |
+
"learning_rate": 9.494098819763953e-05,
|
| 1547 |
+
"loss": 0.764,
|
| 1548 |
+
"step": 25400
|
| 1549 |
+
},
|
| 1550 |
+
{
|
| 1551 |
+
"epoch": 0.51,
|
| 1552 |
+
"learning_rate": 9.492098419683938e-05,
|
| 1553 |
+
"loss": 0.8212,
|
| 1554 |
+
"step": 25500
|
| 1555 |
+
},
|
| 1556 |
+
{
|
| 1557 |
+
"epoch": 0.51,
|
| 1558 |
+
"learning_rate": 9.490098019603922e-05,
|
| 1559 |
+
"loss": 0.7954,
|
| 1560 |
+
"step": 25600
|
| 1561 |
+
},
|
| 1562 |
+
{
|
| 1563 |
+
"epoch": 0.51,
|
| 1564 |
+
"learning_rate": 9.488097619523905e-05,
|
| 1565 |
+
"loss": 0.7965,
|
| 1566 |
+
"step": 25700
|
| 1567 |
+
},
|
| 1568 |
+
{
|
| 1569 |
+
"epoch": 0.52,
|
| 1570 |
+
"learning_rate": 9.48609721944389e-05,
|
| 1571 |
+
"loss": 0.7795,
|
| 1572 |
+
"step": 25800
|
| 1573 |
+
},
|
| 1574 |
+
{
|
| 1575 |
+
"epoch": 0.52,
|
| 1576 |
+
"learning_rate": 9.484096819363874e-05,
|
| 1577 |
+
"loss": 0.7993,
|
| 1578 |
+
"step": 25900
|
| 1579 |
+
},
|
| 1580 |
+
{
|
| 1581 |
+
"epoch": 0.52,
|
| 1582 |
+
"learning_rate": 9.482096419283857e-05,
|
| 1583 |
+
"loss": 0.7951,
|
| 1584 |
+
"step": 26000
|
| 1585 |
+
},
|
| 1586 |
+
{
|
| 1587 |
+
"epoch": 0.52,
|
| 1588 |
+
"learning_rate": 9.480096019203842e-05,
|
| 1589 |
+
"loss": 0.7432,
|
| 1590 |
+
"step": 26100
|
| 1591 |
+
},
|
| 1592 |
+
{
|
| 1593 |
+
"epoch": 0.52,
|
| 1594 |
+
"learning_rate": 9.478095619123826e-05,
|
| 1595 |
+
"loss": 0.7844,
|
| 1596 |
+
"step": 26200
|
| 1597 |
+
},
|
| 1598 |
+
{
|
| 1599 |
+
"epoch": 0.53,
|
| 1600 |
+
"learning_rate": 9.476095219043809e-05,
|
| 1601 |
+
"loss": 0.7029,
|
| 1602 |
+
"step": 26300
|
| 1603 |
+
},
|
| 1604 |
+
{
|
| 1605 |
+
"epoch": 0.53,
|
| 1606 |
+
"learning_rate": 9.474094818963793e-05,
|
| 1607 |
+
"loss": 0.7613,
|
| 1608 |
+
"step": 26400
|
| 1609 |
+
},
|
| 1610 |
+
{
|
| 1611 |
+
"epoch": 0.53,
|
| 1612 |
+
"learning_rate": 9.472094418883777e-05,
|
| 1613 |
+
"loss": 0.7896,
|
| 1614 |
+
"step": 26500
|
| 1615 |
+
},
|
| 1616 |
+
{
|
| 1617 |
+
"epoch": 0.53,
|
| 1618 |
+
"learning_rate": 9.470094018803761e-05,
|
| 1619 |
+
"loss": 0.7674,
|
| 1620 |
+
"step": 26600
|
| 1621 |
+
},
|
| 1622 |
+
{
|
| 1623 |
+
"epoch": 0.53,
|
| 1624 |
+
"learning_rate": 9.468093618723745e-05,
|
| 1625 |
+
"loss": 0.8013,
|
| 1626 |
+
"step": 26700
|
| 1627 |
+
},
|
| 1628 |
+
{
|
| 1629 |
+
"epoch": 0.54,
|
| 1630 |
+
"learning_rate": 9.466093218643729e-05,
|
| 1631 |
+
"loss": 0.811,
|
| 1632 |
+
"step": 26800
|
| 1633 |
+
},
|
| 1634 |
+
{
|
| 1635 |
+
"epoch": 0.54,
|
| 1636 |
+
"learning_rate": 9.464092818563713e-05,
|
| 1637 |
+
"loss": 0.8235,
|
| 1638 |
+
"step": 26900
|
| 1639 |
+
},
|
| 1640 |
+
{
|
| 1641 |
+
"epoch": 0.54,
|
| 1642 |
+
"learning_rate": 9.462092418483696e-05,
|
| 1643 |
+
"loss": 0.7717,
|
| 1644 |
+
"step": 27000
|
| 1645 |
+
},
|
| 1646 |
+
{
|
| 1647 |
+
"epoch": 0.54,
|
| 1648 |
+
"learning_rate": 9.46009201840368e-05,
|
| 1649 |
+
"loss": 0.7364,
|
| 1650 |
+
"step": 27100
|
| 1651 |
+
},
|
| 1652 |
+
{
|
| 1653 |
+
"epoch": 0.54,
|
| 1654 |
+
"learning_rate": 9.458111622324465e-05,
|
| 1655 |
+
"loss": 0.7663,
|
| 1656 |
+
"step": 27200
|
| 1657 |
+
},
|
| 1658 |
+
{
|
| 1659 |
+
"epoch": 0.55,
|
| 1660 |
+
"learning_rate": 9.456111222244449e-05,
|
| 1661 |
+
"loss": 0.771,
|
| 1662 |
+
"step": 27300
|
| 1663 |
+
},
|
| 1664 |
+
{
|
| 1665 |
+
"epoch": 0.55,
|
| 1666 |
+
"learning_rate": 9.454110822164433e-05,
|
| 1667 |
+
"loss": 0.7583,
|
| 1668 |
+
"step": 27400
|
| 1669 |
+
},
|
| 1670 |
+
{
|
| 1671 |
+
"epoch": 0.55,
|
| 1672 |
+
"learning_rate": 9.452110422084417e-05,
|
| 1673 |
+
"loss": 0.7672,
|
| 1674 |
+
"step": 27500
|
| 1675 |
+
},
|
| 1676 |
+
{
|
| 1677 |
+
"epoch": 0.55,
|
| 1678 |
+
"learning_rate": 9.450110022004401e-05,
|
| 1679 |
+
"loss": 0.7706,
|
| 1680 |
+
"step": 27600
|
| 1681 |
+
},
|
| 1682 |
+
{
|
| 1683 |
+
"epoch": 0.55,
|
| 1684 |
+
"learning_rate": 9.448109621924385e-05,
|
| 1685 |
+
"loss": 0.7676,
|
| 1686 |
+
"step": 27700
|
| 1687 |
+
},
|
| 1688 |
+
{
|
| 1689 |
+
"epoch": 0.56,
|
| 1690 |
+
"learning_rate": 9.446109221844369e-05,
|
| 1691 |
+
"loss": 0.7859,
|
| 1692 |
+
"step": 27800
|
| 1693 |
+
},
|
| 1694 |
+
{
|
| 1695 |
+
"epoch": 0.56,
|
| 1696 |
+
"learning_rate": 9.444108821764353e-05,
|
| 1697 |
+
"loss": 0.7573,
|
| 1698 |
+
"step": 27900
|
| 1699 |
+
},
|
| 1700 |
+
{
|
| 1701 |
+
"epoch": 0.56,
|
| 1702 |
+
"learning_rate": 9.442108421684337e-05,
|
| 1703 |
+
"loss": 0.8289,
|
| 1704 |
+
"step": 28000
|
| 1705 |
+
},
|
| 1706 |
+
{
|
| 1707 |
+
"epoch": 0.56,
|
| 1708 |
+
"learning_rate": 9.440108021604322e-05,
|
| 1709 |
+
"loss": 0.7757,
|
| 1710 |
+
"step": 28100
|
| 1711 |
+
},
|
| 1712 |
+
{
|
| 1713 |
+
"epoch": 0.56,
|
| 1714 |
+
"learning_rate": 9.438107621524306e-05,
|
| 1715 |
+
"loss": 0.7688,
|
| 1716 |
+
"step": 28200
|
| 1717 |
+
},
|
| 1718 |
+
{
|
| 1719 |
+
"epoch": 0.57,
|
| 1720 |
+
"learning_rate": 9.43610722144429e-05,
|
| 1721 |
+
"loss": 0.7941,
|
| 1722 |
+
"step": 28300
|
| 1723 |
+
},
|
| 1724 |
+
{
|
| 1725 |
+
"epoch": 0.57,
|
| 1726 |
+
"learning_rate": 9.434106821364274e-05,
|
| 1727 |
+
"loss": 0.7829,
|
| 1728 |
+
"step": 28400
|
| 1729 |
+
},
|
| 1730 |
+
{
|
| 1731 |
+
"epoch": 0.57,
|
| 1732 |
+
"learning_rate": 9.432106421284257e-05,
|
| 1733 |
+
"loss": 0.7744,
|
| 1734 |
+
"step": 28500
|
| 1735 |
+
},
|
| 1736 |
+
{
|
| 1737 |
+
"epoch": 0.57,
|
| 1738 |
+
"learning_rate": 9.430106021204241e-05,
|
| 1739 |
+
"loss": 0.805,
|
| 1740 |
+
"step": 28600
|
| 1741 |
+
},
|
| 1742 |
+
{
|
| 1743 |
+
"epoch": 0.57,
|
| 1744 |
+
"learning_rate": 9.428105621124226e-05,
|
| 1745 |
+
"loss": 0.776,
|
| 1746 |
+
"step": 28700
|
| 1747 |
+
},
|
| 1748 |
+
{
|
| 1749 |
+
"epoch": 0.58,
|
| 1750 |
+
"learning_rate": 9.426105221044209e-05,
|
| 1751 |
+
"loss": 0.7839,
|
| 1752 |
+
"step": 28800
|
| 1753 |
+
},
|
| 1754 |
+
{
|
| 1755 |
+
"epoch": 0.58,
|
| 1756 |
+
"learning_rate": 9.424104820964193e-05,
|
| 1757 |
+
"loss": 0.7894,
|
| 1758 |
+
"step": 28900
|
| 1759 |
+
},
|
| 1760 |
+
{
|
| 1761 |
+
"epoch": 0.58,
|
| 1762 |
+
"learning_rate": 9.422104420884177e-05,
|
| 1763 |
+
"loss": 0.8138,
|
| 1764 |
+
"step": 29000
|
| 1765 |
+
},
|
| 1766 |
+
{
|
| 1767 |
+
"epoch": 0.58,
|
| 1768 |
+
"learning_rate": 9.420104020804161e-05,
|
| 1769 |
+
"loss": 0.7834,
|
| 1770 |
+
"step": 29100
|
| 1771 |
+
},
|
| 1772 |
+
{
|
| 1773 |
+
"epoch": 0.58,
|
| 1774 |
+
"learning_rate": 9.418103620724145e-05,
|
| 1775 |
+
"loss": 0.7323,
|
| 1776 |
+
"step": 29200
|
| 1777 |
+
},
|
| 1778 |
+
{
|
| 1779 |
+
"epoch": 0.59,
|
| 1780 |
+
"learning_rate": 9.416103220644129e-05,
|
| 1781 |
+
"loss": 0.79,
|
| 1782 |
+
"step": 29300
|
| 1783 |
+
},
|
| 1784 |
+
{
|
| 1785 |
+
"epoch": 0.59,
|
| 1786 |
+
"learning_rate": 9.414102820564113e-05,
|
| 1787 |
+
"loss": 0.773,
|
| 1788 |
+
"step": 29400
|
| 1789 |
+
},
|
| 1790 |
+
{
|
| 1791 |
+
"epoch": 0.59,
|
| 1792 |
+
"learning_rate": 9.412102420484097e-05,
|
| 1793 |
+
"loss": 0.7607,
|
| 1794 |
+
"step": 29500
|
| 1795 |
+
},
|
| 1796 |
+
{
|
| 1797 |
+
"epoch": 0.59,
|
| 1798 |
+
"learning_rate": 9.410102020404082e-05,
|
| 1799 |
+
"loss": 0.7581,
|
| 1800 |
+
"step": 29600
|
| 1801 |
+
},
|
| 1802 |
+
{
|
| 1803 |
+
"epoch": 0.59,
|
| 1804 |
+
"learning_rate": 9.408101620324066e-05,
|
| 1805 |
+
"loss": 0.7528,
|
| 1806 |
+
"step": 29700
|
| 1807 |
+
},
|
| 1808 |
+
{
|
| 1809 |
+
"epoch": 0.6,
|
| 1810 |
+
"learning_rate": 9.40610122024405e-05,
|
| 1811 |
+
"loss": 0.7755,
|
| 1812 |
+
"step": 29800
|
| 1813 |
+
},
|
| 1814 |
+
{
|
| 1815 |
+
"epoch": 0.6,
|
| 1816 |
+
"learning_rate": 9.404100820164034e-05,
|
| 1817 |
+
"loss": 0.7633,
|
| 1818 |
+
"step": 29900
|
| 1819 |
+
},
|
| 1820 |
+
{
|
| 1821 |
+
"epoch": 0.6,
|
| 1822 |
+
"learning_rate": 9.402100420084017e-05,
|
| 1823 |
+
"loss": 0.7352,
|
| 1824 |
+
"step": 30000
|
| 1825 |
+
},
|
| 1826 |
+
{
|
| 1827 |
+
"epoch": 0.6,
|
| 1828 |
+
"eval_loss": 0.7202807068824768,
|
| 1829 |
+
"eval_runtime": 707.6198,
|
| 1830 |
+
"eval_samples_per_second": 7.029,
|
| 1831 |
+
"eval_steps_per_second": 0.879,
|
| 1832 |
+
"step": 30000
|
| 1833 |
+
},
|
| 1834 |
+
{
|
| 1835 |
+
"epoch": 0.6,
|
| 1836 |
+
"learning_rate": 9.400100020004001e-05,
|
| 1837 |
+
"loss": 0.7437,
|
| 1838 |
+
"step": 30100
|
| 1839 |
+
},
|
| 1840 |
+
{
|
| 1841 |
+
"epoch": 0.6,
|
| 1842 |
+
"learning_rate": 9.398099619923986e-05,
|
| 1843 |
+
"loss": 0.7887,
|
| 1844 |
+
"step": 30200
|
| 1845 |
+
},
|
| 1846 |
+
{
|
| 1847 |
+
"epoch": 0.61,
|
| 1848 |
+
"learning_rate": 9.396099219843969e-05,
|
| 1849 |
+
"loss": 0.7586,
|
| 1850 |
+
"step": 30300
|
| 1851 |
+
},
|
| 1852 |
+
{
|
| 1853 |
+
"epoch": 0.61,
|
| 1854 |
+
"learning_rate": 9.394098819763953e-05,
|
| 1855 |
+
"loss": 0.7925,
|
| 1856 |
+
"step": 30400
|
| 1857 |
+
},
|
| 1858 |
+
{
|
| 1859 |
+
"epoch": 0.61,
|
| 1860 |
+
"learning_rate": 9.392098419683936e-05,
|
| 1861 |
+
"loss": 0.7947,
|
| 1862 |
+
"step": 30500
|
| 1863 |
+
},
|
| 1864 |
+
{
|
| 1865 |
+
"epoch": 0.61,
|
| 1866 |
+
"learning_rate": 9.390098019603921e-05,
|
| 1867 |
+
"loss": 0.7701,
|
| 1868 |
+
"step": 30600
|
| 1869 |
+
},
|
| 1870 |
+
{
|
| 1871 |
+
"epoch": 0.61,
|
| 1872 |
+
"learning_rate": 9.388117623524706e-05,
|
| 1873 |
+
"loss": 0.7631,
|
| 1874 |
+
"step": 30700
|
| 1875 |
+
},
|
| 1876 |
+
{
|
| 1877 |
+
"epoch": 0.62,
|
| 1878 |
+
"learning_rate": 9.38611722344469e-05,
|
| 1879 |
+
"loss": 0.7977,
|
| 1880 |
+
"step": 30800
|
| 1881 |
+
},
|
| 1882 |
+
{
|
| 1883 |
+
"epoch": 0.62,
|
| 1884 |
+
"learning_rate": 9.384116823364674e-05,
|
| 1885 |
+
"loss": 0.7941,
|
| 1886 |
+
"step": 30900
|
| 1887 |
+
},
|
| 1888 |
+
{
|
| 1889 |
+
"epoch": 0.62,
|
| 1890 |
+
"learning_rate": 9.382116423284657e-05,
|
| 1891 |
+
"loss": 0.7929,
|
| 1892 |
+
"step": 31000
|
| 1893 |
+
},
|
| 1894 |
+
{
|
| 1895 |
+
"epoch": 0.62,
|
| 1896 |
+
"learning_rate": 9.380116023204641e-05,
|
| 1897 |
+
"loss": 0.7871,
|
| 1898 |
+
"step": 31100
|
| 1899 |
+
},
|
| 1900 |
+
{
|
| 1901 |
+
"epoch": 0.62,
|
| 1902 |
+
"learning_rate": 9.378115623124626e-05,
|
| 1903 |
+
"loss": 0.7449,
|
| 1904 |
+
"step": 31200
|
| 1905 |
+
},
|
| 1906 |
+
{
|
| 1907 |
+
"epoch": 0.63,
|
| 1908 |
+
"learning_rate": 9.376115223044609e-05,
|
| 1909 |
+
"loss": 0.7994,
|
| 1910 |
+
"step": 31300
|
| 1911 |
+
},
|
| 1912 |
+
{
|
| 1913 |
+
"epoch": 0.63,
|
| 1914 |
+
"learning_rate": 9.374114822964593e-05,
|
| 1915 |
+
"loss": 0.8307,
|
| 1916 |
+
"step": 31400
|
| 1917 |
+
},
|
| 1918 |
+
{
|
| 1919 |
+
"epoch": 0.63,
|
| 1920 |
+
"learning_rate": 9.372114422884576e-05,
|
| 1921 |
+
"loss": 0.7521,
|
| 1922 |
+
"step": 31500
|
| 1923 |
+
},
|
| 1924 |
+
{
|
| 1925 |
+
"epoch": 0.63,
|
| 1926 |
+
"learning_rate": 9.370114022804561e-05,
|
| 1927 |
+
"loss": 0.7767,
|
| 1928 |
+
"step": 31600
|
| 1929 |
+
},
|
| 1930 |
+
{
|
| 1931 |
+
"epoch": 0.63,
|
| 1932 |
+
"learning_rate": 9.368113622724545e-05,
|
| 1933 |
+
"loss": 0.7686,
|
| 1934 |
+
"step": 31700
|
| 1935 |
+
},
|
| 1936 |
+
{
|
| 1937 |
+
"epoch": 0.64,
|
| 1938 |
+
"learning_rate": 9.36611322264453e-05,
|
| 1939 |
+
"loss": 0.7903,
|
| 1940 |
+
"step": 31800
|
| 1941 |
+
},
|
| 1942 |
+
{
|
| 1943 |
+
"epoch": 0.64,
|
| 1944 |
+
"learning_rate": 9.364112822564514e-05,
|
| 1945 |
+
"loss": 0.7247,
|
| 1946 |
+
"step": 31900
|
| 1947 |
+
},
|
| 1948 |
+
{
|
| 1949 |
+
"epoch": 0.64,
|
| 1950 |
+
"learning_rate": 9.362112422484497e-05,
|
| 1951 |
+
"loss": 0.7425,
|
| 1952 |
+
"step": 32000
|
| 1953 |
+
},
|
| 1954 |
+
{
|
| 1955 |
+
"epoch": 0.64,
|
| 1956 |
+
"learning_rate": 9.360112022404482e-05,
|
| 1957 |
+
"loss": 0.7607,
|
| 1958 |
+
"step": 32100
|
| 1959 |
+
},
|
| 1960 |
+
{
|
| 1961 |
+
"epoch": 0.64,
|
| 1962 |
+
"learning_rate": 9.358111622324466e-05,
|
| 1963 |
+
"loss": 0.7745,
|
| 1964 |
+
"step": 32200
|
| 1965 |
+
},
|
| 1966 |
+
{
|
| 1967 |
+
"epoch": 0.65,
|
| 1968 |
+
"learning_rate": 9.356111222244449e-05,
|
| 1969 |
+
"loss": 0.7873,
|
| 1970 |
+
"step": 32300
|
| 1971 |
+
},
|
| 1972 |
+
{
|
| 1973 |
+
"epoch": 0.65,
|
| 1974 |
+
"learning_rate": 9.354110822164434e-05,
|
| 1975 |
+
"loss": 0.7496,
|
| 1976 |
+
"step": 32400
|
| 1977 |
+
},
|
| 1978 |
+
{
|
| 1979 |
+
"epoch": 0.65,
|
| 1980 |
+
"learning_rate": 9.352110422084417e-05,
|
| 1981 |
+
"loss": 0.7789,
|
| 1982 |
+
"step": 32500
|
| 1983 |
+
},
|
| 1984 |
+
{
|
| 1985 |
+
"epoch": 0.65,
|
| 1986 |
+
"learning_rate": 9.350110022004401e-05,
|
| 1987 |
+
"loss": 0.7862,
|
| 1988 |
+
"step": 32600
|
| 1989 |
+
},
|
| 1990 |
+
{
|
| 1991 |
+
"epoch": 0.65,
|
| 1992 |
+
"learning_rate": 9.348109621924386e-05,
|
| 1993 |
+
"loss": 0.7496,
|
| 1994 |
+
"step": 32700
|
| 1995 |
+
},
|
| 1996 |
+
{
|
| 1997 |
+
"epoch": 0.66,
|
| 1998 |
+
"learning_rate": 9.346109221844369e-05,
|
| 1999 |
+
"loss": 0.7601,
|
| 2000 |
+
"step": 32800
|
| 2001 |
+
},
|
| 2002 |
+
{
|
| 2003 |
+
"epoch": 0.66,
|
| 2004 |
+
"learning_rate": 9.344108821764353e-05,
|
| 2005 |
+
"loss": 0.822,
|
| 2006 |
+
"step": 32900
|
| 2007 |
+
},
|
| 2008 |
+
{
|
| 2009 |
+
"epoch": 0.66,
|
| 2010 |
+
"learning_rate": 9.342108421684336e-05,
|
| 2011 |
+
"loss": 0.7935,
|
| 2012 |
+
"step": 33000
|
| 2013 |
+
},
|
| 2014 |
+
{
|
| 2015 |
+
"epoch": 0.66,
|
| 2016 |
+
"learning_rate": 9.340108021604321e-05,
|
| 2017 |
+
"loss": 0.7496,
|
| 2018 |
+
"step": 33100
|
| 2019 |
+
},
|
| 2020 |
+
{
|
| 2021 |
+
"epoch": 0.66,
|
| 2022 |
+
"learning_rate": 9.338107621524305e-05,
|
| 2023 |
+
"loss": 0.7517,
|
| 2024 |
+
"step": 33200
|
| 2025 |
+
},
|
| 2026 |
+
{
|
| 2027 |
+
"epoch": 0.67,
|
| 2028 |
+
"learning_rate": 9.33610722144429e-05,
|
| 2029 |
+
"loss": 0.7648,
|
| 2030 |
+
"step": 33300
|
| 2031 |
+
},
|
| 2032 |
+
{
|
| 2033 |
+
"epoch": 0.67,
|
| 2034 |
+
"learning_rate": 9.334106821364274e-05,
|
| 2035 |
+
"loss": 0.7855,
|
| 2036 |
+
"step": 33400
|
| 2037 |
+
},
|
| 2038 |
+
{
|
| 2039 |
+
"epoch": 0.67,
|
| 2040 |
+
"learning_rate": 9.332106421284257e-05,
|
| 2041 |
+
"loss": 0.8075,
|
| 2042 |
+
"step": 33500
|
| 2043 |
+
},
|
| 2044 |
+
{
|
| 2045 |
+
"epoch": 0.67,
|
| 2046 |
+
"learning_rate": 9.330106021204242e-05,
|
| 2047 |
+
"loss": 0.7918,
|
| 2048 |
+
"step": 33600
|
| 2049 |
+
},
|
| 2050 |
+
{
|
| 2051 |
+
"epoch": 0.67,
|
| 2052 |
+
"learning_rate": 9.328105621124226e-05,
|
| 2053 |
+
"loss": 0.7654,
|
| 2054 |
+
"step": 33700
|
| 2055 |
+
},
|
| 2056 |
+
{
|
| 2057 |
+
"epoch": 0.68,
|
| 2058 |
+
"learning_rate": 9.326105221044209e-05,
|
| 2059 |
+
"loss": 0.7865,
|
| 2060 |
+
"step": 33800
|
| 2061 |
+
},
|
| 2062 |
+
{
|
| 2063 |
+
"epoch": 0.68,
|
| 2064 |
+
"learning_rate": 9.324104820964194e-05,
|
| 2065 |
+
"loss": 0.7665,
|
| 2066 |
+
"step": 33900
|
| 2067 |
+
},
|
| 2068 |
+
{
|
| 2069 |
+
"epoch": 0.68,
|
| 2070 |
+
"learning_rate": 9.322104420884177e-05,
|
| 2071 |
+
"loss": 0.7662,
|
| 2072 |
+
"step": 34000
|
| 2073 |
+
},
|
| 2074 |
+
{
|
| 2075 |
+
"epoch": 0.68,
|
| 2076 |
+
"learning_rate": 9.320104020804161e-05,
|
| 2077 |
+
"loss": 0.7713,
|
| 2078 |
+
"step": 34100
|
| 2079 |
+
},
|
| 2080 |
+
{
|
| 2081 |
+
"epoch": 0.68,
|
| 2082 |
+
"learning_rate": 9.318103620724146e-05,
|
| 2083 |
+
"loss": 0.7528,
|
| 2084 |
+
"step": 34200
|
| 2085 |
+
},
|
| 2086 |
+
{
|
| 2087 |
+
"epoch": 0.69,
|
| 2088 |
+
"learning_rate": 9.316103220644129e-05,
|
| 2089 |
+
"loss": 0.7992,
|
| 2090 |
+
"step": 34300
|
| 2091 |
+
},
|
| 2092 |
+
{
|
| 2093 |
+
"epoch": 0.69,
|
| 2094 |
+
"learning_rate": 9.314102820564113e-05,
|
| 2095 |
+
"loss": 0.762,
|
| 2096 |
+
"step": 34400
|
| 2097 |
+
},
|
| 2098 |
+
{
|
| 2099 |
+
"epoch": 0.69,
|
| 2100 |
+
"learning_rate": 9.312122424484897e-05,
|
| 2101 |
+
"loss": 0.7782,
|
| 2102 |
+
"step": 34500
|
| 2103 |
+
},
|
| 2104 |
+
{
|
| 2105 |
+
"epoch": 0.69,
|
| 2106 |
+
"learning_rate": 9.310122024404882e-05,
|
| 2107 |
+
"loss": 0.764,
|
| 2108 |
+
"step": 34600
|
| 2109 |
+
},
|
| 2110 |
+
{
|
| 2111 |
+
"epoch": 0.69,
|
| 2112 |
+
"learning_rate": 9.308121624324866e-05,
|
| 2113 |
+
"loss": 0.7462,
|
| 2114 |
+
"step": 34700
|
| 2115 |
+
},
|
| 2116 |
+
{
|
| 2117 |
+
"epoch": 0.7,
|
| 2118 |
+
"learning_rate": 9.306121224244849e-05,
|
| 2119 |
+
"loss": 0.7737,
|
| 2120 |
+
"step": 34800
|
| 2121 |
+
},
|
| 2122 |
+
{
|
| 2123 |
+
"epoch": 0.7,
|
| 2124 |
+
"learning_rate": 9.304120824164834e-05,
|
| 2125 |
+
"loss": 0.7289,
|
| 2126 |
+
"step": 34900
|
| 2127 |
+
},
|
| 2128 |
+
{
|
| 2129 |
+
"epoch": 0.7,
|
| 2130 |
+
"learning_rate": 9.302120424084817e-05,
|
| 2131 |
+
"loss": 0.7715,
|
| 2132 |
+
"step": 35000
|
| 2133 |
+
},
|
| 2134 |
+
{
|
| 2135 |
+
"epoch": 0.7,
|
| 2136 |
+
"learning_rate": 9.300120024004801e-05,
|
| 2137 |
+
"loss": 0.7211,
|
| 2138 |
+
"step": 35100
|
| 2139 |
+
},
|
| 2140 |
+
{
|
| 2141 |
+
"epoch": 0.7,
|
| 2142 |
+
"learning_rate": 9.298119623924786e-05,
|
| 2143 |
+
"loss": 0.7473,
|
| 2144 |
+
"step": 35200
|
| 2145 |
+
},
|
| 2146 |
+
{
|
| 2147 |
+
"epoch": 0.71,
|
| 2148 |
+
"learning_rate": 9.296119223844769e-05,
|
| 2149 |
+
"loss": 0.7805,
|
| 2150 |
+
"step": 35300
|
| 2151 |
+
},
|
| 2152 |
+
{
|
| 2153 |
+
"epoch": 0.71,
|
| 2154 |
+
"learning_rate": 9.294118823764753e-05,
|
| 2155 |
+
"loss": 0.7617,
|
| 2156 |
+
"step": 35400
|
| 2157 |
+
},
|
| 2158 |
+
{
|
| 2159 |
+
"epoch": 0.71,
|
| 2160 |
+
"learning_rate": 9.292118423684736e-05,
|
| 2161 |
+
"loss": 0.7964,
|
| 2162 |
+
"step": 35500
|
| 2163 |
+
},
|
| 2164 |
+
{
|
| 2165 |
+
"epoch": 0.71,
|
| 2166 |
+
"learning_rate": 9.290118023604721e-05,
|
| 2167 |
+
"loss": 0.7889,
|
| 2168 |
+
"step": 35600
|
| 2169 |
+
},
|
| 2170 |
+
{
|
| 2171 |
+
"epoch": 0.71,
|
| 2172 |
+
"learning_rate": 9.288117623524705e-05,
|
| 2173 |
+
"loss": 0.7584,
|
| 2174 |
+
"step": 35700
|
| 2175 |
+
},
|
| 2176 |
+
{
|
| 2177 |
+
"epoch": 0.72,
|
| 2178 |
+
"learning_rate": 9.28611722344469e-05,
|
| 2179 |
+
"loss": 0.8073,
|
| 2180 |
+
"step": 35800
|
| 2181 |
+
},
|
| 2182 |
+
{
|
| 2183 |
+
"epoch": 0.72,
|
| 2184 |
+
"learning_rate": 9.284116823364674e-05,
|
| 2185 |
+
"loss": 0.7515,
|
| 2186 |
+
"step": 35900
|
| 2187 |
+
},
|
| 2188 |
+
{
|
| 2189 |
+
"epoch": 0.72,
|
| 2190 |
+
"learning_rate": 9.282116423284657e-05,
|
| 2191 |
+
"loss": 0.7605,
|
| 2192 |
+
"step": 36000
|
| 2193 |
+
},
|
| 2194 |
+
{
|
| 2195 |
+
"epoch": 0.72,
|
| 2196 |
+
"learning_rate": 9.280116023204642e-05,
|
| 2197 |
+
"loss": 0.7869,
|
| 2198 |
+
"step": 36100
|
| 2199 |
+
},
|
| 2200 |
+
{
|
| 2201 |
+
"epoch": 0.72,
|
| 2202 |
+
"learning_rate": 9.278115623124626e-05,
|
| 2203 |
+
"loss": 0.7255,
|
| 2204 |
+
"step": 36200
|
| 2205 |
+
},
|
| 2206 |
+
{
|
| 2207 |
+
"epoch": 0.73,
|
| 2208 |
+
"learning_rate": 9.276115223044609e-05,
|
| 2209 |
+
"loss": 0.7429,
|
| 2210 |
+
"step": 36300
|
| 2211 |
+
},
|
| 2212 |
+
{
|
| 2213 |
+
"epoch": 0.73,
|
| 2214 |
+
"learning_rate": 9.274114822964594e-05,
|
| 2215 |
+
"loss": 0.7826,
|
| 2216 |
+
"step": 36400
|
| 2217 |
+
},
|
| 2218 |
+
{
|
| 2219 |
+
"epoch": 0.73,
|
| 2220 |
+
"learning_rate": 9.272114422884577e-05,
|
| 2221 |
+
"loss": 0.7448,
|
| 2222 |
+
"step": 36500
|
| 2223 |
+
},
|
| 2224 |
+
{
|
| 2225 |
+
"epoch": 0.73,
|
| 2226 |
+
"learning_rate": 9.270114022804561e-05,
|
| 2227 |
+
"loss": 0.7862,
|
| 2228 |
+
"step": 36600
|
| 2229 |
+
},
|
| 2230 |
+
{
|
| 2231 |
+
"epoch": 0.73,
|
| 2232 |
+
"learning_rate": 9.268113622724546e-05,
|
| 2233 |
+
"loss": 0.7566,
|
| 2234 |
+
"step": 36700
|
| 2235 |
+
},
|
| 2236 |
+
{
|
| 2237 |
+
"epoch": 0.74,
|
| 2238 |
+
"learning_rate": 9.266113222644529e-05,
|
| 2239 |
+
"loss": 0.7131,
|
| 2240 |
+
"step": 36800
|
| 2241 |
+
},
|
| 2242 |
+
{
|
| 2243 |
+
"epoch": 0.74,
|
| 2244 |
+
"learning_rate": 9.264112822564513e-05,
|
| 2245 |
+
"loss": 0.77,
|
| 2246 |
+
"step": 36900
|
| 2247 |
+
},
|
| 2248 |
+
{
|
| 2249 |
+
"epoch": 0.74,
|
| 2250 |
+
"learning_rate": 9.262112422484496e-05,
|
| 2251 |
+
"loss": 0.7562,
|
| 2252 |
+
"step": 37000
|
| 2253 |
+
},
|
| 2254 |
+
{
|
| 2255 |
+
"epoch": 0.74,
|
| 2256 |
+
"learning_rate": 9.260112022404481e-05,
|
| 2257 |
+
"loss": 0.7475,
|
| 2258 |
+
"step": 37100
|
| 2259 |
+
},
|
| 2260 |
+
{
|
| 2261 |
+
"epoch": 0.74,
|
| 2262 |
+
"learning_rate": 9.258111622324465e-05,
|
| 2263 |
+
"loss": 0.7773,
|
| 2264 |
+
"step": 37200
|
| 2265 |
+
},
|
| 2266 |
+
{
|
| 2267 |
+
"epoch": 0.75,
|
| 2268 |
+
"learning_rate": 9.25611122224445e-05,
|
| 2269 |
+
"loss": 0.7671,
|
| 2270 |
+
"step": 37300
|
| 2271 |
+
},
|
| 2272 |
+
{
|
| 2273 |
+
"epoch": 0.75,
|
| 2274 |
+
"learning_rate": 9.254130826165234e-05,
|
| 2275 |
+
"loss": 0.7903,
|
| 2276 |
+
"step": 37400
|
| 2277 |
+
},
|
| 2278 |
+
{
|
| 2279 |
+
"epoch": 0.75,
|
| 2280 |
+
"learning_rate": 9.252130426085217e-05,
|
| 2281 |
+
"loss": 0.8081,
|
| 2282 |
+
"step": 37500
|
| 2283 |
+
},
|
| 2284 |
+
{
|
| 2285 |
+
"epoch": 0.75,
|
| 2286 |
+
"learning_rate": 9.250130026005201e-05,
|
| 2287 |
+
"loss": 0.7324,
|
| 2288 |
+
"step": 37600
|
| 2289 |
+
},
|
| 2290 |
+
{
|
| 2291 |
+
"epoch": 0.75,
|
| 2292 |
+
"learning_rate": 9.248129625925186e-05,
|
| 2293 |
+
"loss": 0.7595,
|
| 2294 |
+
"step": 37700
|
| 2295 |
+
},
|
| 2296 |
+
{
|
| 2297 |
+
"epoch": 0.76,
|
| 2298 |
+
"learning_rate": 9.246129225845169e-05,
|
| 2299 |
+
"loss": 0.774,
|
| 2300 |
+
"step": 37800
|
| 2301 |
+
},
|
| 2302 |
+
{
|
| 2303 |
+
"epoch": 0.76,
|
| 2304 |
+
"learning_rate": 9.244128825765153e-05,
|
| 2305 |
+
"loss": 0.8066,
|
| 2306 |
+
"step": 37900
|
| 2307 |
+
},
|
| 2308 |
+
{
|
| 2309 |
+
"epoch": 0.76,
|
| 2310 |
+
"learning_rate": 9.242128425685138e-05,
|
| 2311 |
+
"loss": 0.7878,
|
| 2312 |
+
"step": 38000
|
| 2313 |
+
},
|
| 2314 |
+
{
|
| 2315 |
+
"epoch": 0.76,
|
| 2316 |
+
"learning_rate": 9.240128025605122e-05,
|
| 2317 |
+
"loss": 0.781,
|
| 2318 |
+
"step": 38100
|
| 2319 |
+
},
|
| 2320 |
+
{
|
| 2321 |
+
"epoch": 0.76,
|
| 2322 |
+
"learning_rate": 9.238127625525106e-05,
|
| 2323 |
+
"loss": 0.733,
|
| 2324 |
+
"step": 38200
|
| 2325 |
+
},
|
| 2326 |
+
{
|
| 2327 |
+
"epoch": 0.77,
|
| 2328 |
+
"learning_rate": 9.23612722544509e-05,
|
| 2329 |
+
"loss": 0.7704,
|
| 2330 |
+
"step": 38300
|
| 2331 |
+
},
|
| 2332 |
+
{
|
| 2333 |
+
"epoch": 0.77,
|
| 2334 |
+
"learning_rate": 9.234126825365074e-05,
|
| 2335 |
+
"loss": 0.7342,
|
| 2336 |
+
"step": 38400
|
| 2337 |
+
},
|
| 2338 |
+
{
|
| 2339 |
+
"epoch": 0.77,
|
| 2340 |
+
"learning_rate": 9.232126425285057e-05,
|
| 2341 |
+
"loss": 0.7681,
|
| 2342 |
+
"step": 38500
|
| 2343 |
+
},
|
| 2344 |
+
{
|
| 2345 |
+
"epoch": 0.77,
|
| 2346 |
+
"learning_rate": 9.230126025205042e-05,
|
| 2347 |
+
"loss": 0.7074,
|
| 2348 |
+
"step": 38600
|
| 2349 |
+
},
|
| 2350 |
+
{
|
| 2351 |
+
"epoch": 0.77,
|
| 2352 |
+
"learning_rate": 9.228125625125026e-05,
|
| 2353 |
+
"loss": 0.7909,
|
| 2354 |
+
"step": 38700
|
| 2355 |
+
},
|
| 2356 |
+
{
|
| 2357 |
+
"epoch": 0.78,
|
| 2358 |
+
"learning_rate": 9.226125225045009e-05,
|
| 2359 |
+
"loss": 0.7581,
|
| 2360 |
+
"step": 38800
|
| 2361 |
+
},
|
| 2362 |
+
{
|
| 2363 |
+
"epoch": 0.78,
|
| 2364 |
+
"learning_rate": 9.224124824964994e-05,
|
| 2365 |
+
"loss": 0.7346,
|
| 2366 |
+
"step": 38900
|
| 2367 |
+
},
|
| 2368 |
+
{
|
| 2369 |
+
"epoch": 0.78,
|
| 2370 |
+
"learning_rate": 9.222124424884977e-05,
|
| 2371 |
+
"loss": 0.7564,
|
| 2372 |
+
"step": 39000
|
| 2373 |
+
},
|
| 2374 |
+
{
|
| 2375 |
+
"epoch": 0.78,
|
| 2376 |
+
"learning_rate": 9.220124024804961e-05,
|
| 2377 |
+
"loss": 0.8259,
|
| 2378 |
+
"step": 39100
|
| 2379 |
+
},
|
| 2380 |
+
{
|
| 2381 |
+
"epoch": 0.78,
|
| 2382 |
+
"learning_rate": 9.218123624724946e-05,
|
| 2383 |
+
"loss": 0.7828,
|
| 2384 |
+
"step": 39200
|
| 2385 |
+
},
|
| 2386 |
+
{
|
| 2387 |
+
"epoch": 0.79,
|
| 2388 |
+
"learning_rate": 9.216123224644929e-05,
|
| 2389 |
+
"loss": 0.7757,
|
| 2390 |
+
"step": 39300
|
| 2391 |
+
},
|
| 2392 |
+
{
|
| 2393 |
+
"epoch": 0.79,
|
| 2394 |
+
"learning_rate": 9.214122824564913e-05,
|
| 2395 |
+
"loss": 0.7402,
|
| 2396 |
+
"step": 39400
|
| 2397 |
+
},
|
| 2398 |
+
{
|
| 2399 |
+
"epoch": 0.79,
|
| 2400 |
+
"learning_rate": 9.212122424484898e-05,
|
| 2401 |
+
"loss": 0.807,
|
| 2402 |
+
"step": 39500
|
| 2403 |
+
},
|
| 2404 |
+
{
|
| 2405 |
+
"epoch": 0.79,
|
| 2406 |
+
"learning_rate": 9.210122024404882e-05,
|
| 2407 |
+
"loss": 0.7574,
|
| 2408 |
+
"step": 39600
|
| 2409 |
+
},
|
| 2410 |
+
{
|
| 2411 |
+
"epoch": 0.79,
|
| 2412 |
+
"learning_rate": 9.208121624324866e-05,
|
| 2413 |
+
"loss": 0.7471,
|
| 2414 |
+
"step": 39700
|
| 2415 |
+
},
|
| 2416 |
+
{
|
| 2417 |
+
"epoch": 0.8,
|
| 2418 |
+
"learning_rate": 9.20612122424485e-05,
|
| 2419 |
+
"loss": 0.7779,
|
| 2420 |
+
"step": 39800
|
| 2421 |
+
},
|
| 2422 |
+
{
|
| 2423 |
+
"epoch": 0.8,
|
| 2424 |
+
"learning_rate": 9.204140828165634e-05,
|
| 2425 |
+
"loss": 0.7553,
|
| 2426 |
+
"step": 39900
|
| 2427 |
+
},
|
| 2428 |
+
{
|
| 2429 |
+
"epoch": 0.8,
|
| 2430 |
+
"learning_rate": 9.202140428085617e-05,
|
| 2431 |
+
"loss": 0.796,
|
| 2432 |
+
"step": 40000
|
| 2433 |
+
},
|
| 2434 |
+
{
|
| 2435 |
+
"epoch": 0.8,
|
| 2436 |
+
"eval_loss": 0.7169719338417053,
|
| 2437 |
+
"eval_runtime": 690.7702,
|
| 2438 |
+
"eval_samples_per_second": 7.201,
|
| 2439 |
+
"eval_steps_per_second": 0.9,
|
| 2440 |
+
"step": 40000
|
| 2441 |
+
}
|
| 2442 |
+
],
|
| 2443 |
+
"max_steps": 500000,
|
| 2444 |
+
"num_train_epochs": 10,
|
| 2445 |
+
"total_flos": 5.6288641646709965e+17,
|
| 2446 |
+
"trial_name": null,
|
| 2447 |
+
"trial_params": null
|
| 2448 |
+
}
|
checkpoint-40000/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:86e3e950534e1e3e08a50aef6d282bea64ab70d530ec27d5c1bcac285ba164ae
|
| 3 |
+
size 4408
|