Upload 12 files
Browse files- .gitattributes +1 -0
- config.json +30 -0
- generation_config.json +6 -0
- model.safetensors +3 -0
- optimizer.pt +3 -0
- rng_state.pth +3 -0
- scheduler.pt +3 -0
- special_tokens_map.json +0 -0
- tokenizer.json +3 -0
- tokenizer_config.json +0 -0
- trainer_state.json +1083 -0
- training_args.bin +3 -0
- vocab.json +0 -0
.gitattributes
CHANGED
|
@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
|
|
| 33 |
*.zip filter=lfs diff=lfs merge=lfs -text
|
| 34 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
| 35 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
|
|
|
|
|
| 33 |
*.zip filter=lfs diff=lfs merge=lfs -text
|
| 34 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
| 35 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
| 36 |
+
tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
config.json
ADDED
|
@@ -0,0 +1,30 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "D:\\AI_APPS\\OuteTTS\\OuteTTS-0.3-500M",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"Qwen2ForCausalLM"
|
| 5 |
+
],
|
| 6 |
+
"attention_dropout": 0.0,
|
| 7 |
+
"bos_token_id": 151643,
|
| 8 |
+
"eos_token_id": 151645,
|
| 9 |
+
"hidden_act": "silu",
|
| 10 |
+
"hidden_size": 896,
|
| 11 |
+
"initializer_range": 0.02,
|
| 12 |
+
"intermediate_size": 4864,
|
| 13 |
+
"max_position_embeddings": 32768,
|
| 14 |
+
"max_window_layers": 24,
|
| 15 |
+
"model_type": "qwen2",
|
| 16 |
+
"num_attention_heads": 14,
|
| 17 |
+
"num_hidden_layers": 24,
|
| 18 |
+
"num_key_value_heads": 2,
|
| 19 |
+
"rms_norm_eps": 1e-06,
|
| 20 |
+
"rope_scaling": null,
|
| 21 |
+
"rope_theta": 1000000.0,
|
| 22 |
+
"sliding_window": null,
|
| 23 |
+
"tie_word_embeddings": true,
|
| 24 |
+
"torch_dtype": "float32",
|
| 25 |
+
"transformers_version": "4.48.3",
|
| 26 |
+
"use_cache": true,
|
| 27 |
+
"use_mrope": false,
|
| 28 |
+
"use_sliding_window": false,
|
| 29 |
+
"vocab_size": 157696
|
| 30 |
+
}
|
generation_config.json
ADDED
|
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token_id": 151643,
|
| 3 |
+
"eos_token_id": 151645,
|
| 4 |
+
"pad_token_id": 151645,
|
| 5 |
+
"transformers_version": "4.48.3"
|
| 6 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:440718d6f157a98fd22fc60f869ae16440cb4fe7003c91d7ecea9497a6e2ae5c
|
| 3 |
+
size 1996807320
|
optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:603296525381dbef5032aba1f1295f5eb51f183d81a3b67d62c15d625b45a048
|
| 3 |
+
size 3993792954
|
rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:cdf4d700cdd7ef81bf005d2842e2f669e1810471cfce670ab79872c6ee698af7
|
| 3 |
+
size 14244
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a49e29b758d40cfe1c153cb8b95a52af827c195e13df13a4a897c5cb79216700
|
| 3 |
+
size 1064
|
special_tokens_map.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tokenizer.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9dd411277ccf5ecddf0cb0c7133a0f9be0c5a5d5e9f20874a1f10512429fbcd7
|
| 3 |
+
size 12371875
|
tokenizer_config.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
trainer_state.json
ADDED
|
@@ -0,0 +1,1083 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.03161089100564781,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 1500,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.00021073927337098542,
|
| 13 |
+
"grad_norm": 4.289033889770508,
|
| 14 |
+
"learning_rate": 4.998946303633145e-05,
|
| 15 |
+
"loss": 3.8267,
|
| 16 |
+
"step": 10
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.00042147854674197085,
|
| 20 |
+
"grad_norm": 4.060084342956543,
|
| 21 |
+
"learning_rate": 4.997997976902976e-05,
|
| 22 |
+
"loss": 3.6144,
|
| 23 |
+
"step": 20
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.0006322178201129563,
|
| 27 |
+
"grad_norm": 4.173689842224121,
|
| 28 |
+
"learning_rate": 4.996944280536121e-05,
|
| 29 |
+
"loss": 3.4868,
|
| 30 |
+
"step": 30
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.0008429570934839417,
|
| 34 |
+
"grad_norm": 5.22714376449585,
|
| 35 |
+
"learning_rate": 4.995890584169266e-05,
|
| 36 |
+
"loss": 3.4488,
|
| 37 |
+
"step": 40
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.001053696366854927,
|
| 41 |
+
"grad_norm": 5.710427284240723,
|
| 42 |
+
"learning_rate": 4.994836887802411e-05,
|
| 43 |
+
"loss": 3.3072,
|
| 44 |
+
"step": 50
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.0012644356402259126,
|
| 48 |
+
"grad_norm": 5.226071357727051,
|
| 49 |
+
"learning_rate": 4.993783191435556e-05,
|
| 50 |
+
"loss": 3.3811,
|
| 51 |
+
"step": 60
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.0014751749135968979,
|
| 55 |
+
"grad_norm": 4.471126556396484,
|
| 56 |
+
"learning_rate": 4.9927294950687014e-05,
|
| 57 |
+
"loss": 3.3294,
|
| 58 |
+
"step": 70
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.0016859141869678834,
|
| 62 |
+
"grad_norm": 3.511232376098633,
|
| 63 |
+
"learning_rate": 4.991675798701846e-05,
|
| 64 |
+
"loss": 3.274,
|
| 65 |
+
"step": 80
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.0018966534603388687,
|
| 69 |
+
"grad_norm": 3.6000168323516846,
|
| 70 |
+
"learning_rate": 4.990622102334991e-05,
|
| 71 |
+
"loss": 3.2927,
|
| 72 |
+
"step": 90
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.002107392733709854,
|
| 76 |
+
"grad_norm": 3.6504461765289307,
|
| 77 |
+
"learning_rate": 4.9895684059681366e-05,
|
| 78 |
+
"loss": 3.1517,
|
| 79 |
+
"step": 100
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.0023181320070808395,
|
| 83 |
+
"grad_norm": 5.8762288093566895,
|
| 84 |
+
"learning_rate": 4.9885147096012814e-05,
|
| 85 |
+
"loss": 3.2165,
|
| 86 |
+
"step": 110
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.002528871280451825,
|
| 90 |
+
"grad_norm": 3.862133264541626,
|
| 91 |
+
"learning_rate": 4.987461013234427e-05,
|
| 92 |
+
"loss": 3.2496,
|
| 93 |
+
"step": 120
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.0027396105538228105,
|
| 97 |
+
"grad_norm": 3.1009256839752197,
|
| 98 |
+
"learning_rate": 4.986407316867572e-05,
|
| 99 |
+
"loss": 3.1269,
|
| 100 |
+
"step": 130
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.0029503498271937958,
|
| 104 |
+
"grad_norm": 4.951446056365967,
|
| 105 |
+
"learning_rate": 4.9853536205007165e-05,
|
| 106 |
+
"loss": 3.142,
|
| 107 |
+
"step": 140
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.003161089100564781,
|
| 111 |
+
"grad_norm": 4.305763244628906,
|
| 112 |
+
"learning_rate": 4.9842999241338614e-05,
|
| 113 |
+
"loss": 3.1518,
|
| 114 |
+
"step": 150
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.003371828373935767,
|
| 118 |
+
"grad_norm": 4.36134672164917,
|
| 119 |
+
"learning_rate": 4.983246227767007e-05,
|
| 120 |
+
"loss": 3.1285,
|
| 121 |
+
"step": 160
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.003582567647306752,
|
| 125 |
+
"grad_norm": 3.9053471088409424,
|
| 126 |
+
"learning_rate": 4.9821925314001524e-05,
|
| 127 |
+
"loss": 3.0738,
|
| 128 |
+
"step": 170
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.0037933069206777374,
|
| 132 |
+
"grad_norm": 13.242171287536621,
|
| 133 |
+
"learning_rate": 4.981138835033297e-05,
|
| 134 |
+
"loss": 3.0425,
|
| 135 |
+
"step": 180
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.004004046194048723,
|
| 139 |
+
"grad_norm": 3.8637614250183105,
|
| 140 |
+
"learning_rate": 4.980085138666442e-05,
|
| 141 |
+
"loss": 3.007,
|
| 142 |
+
"step": 190
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.004214785467419708,
|
| 146 |
+
"grad_norm": 4.488512992858887,
|
| 147 |
+
"learning_rate": 4.979031442299587e-05,
|
| 148 |
+
"loss": 3.0578,
|
| 149 |
+
"step": 200
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.004425524740790694,
|
| 153 |
+
"grad_norm": 5.628246307373047,
|
| 154 |
+
"learning_rate": 4.9779777459327324e-05,
|
| 155 |
+
"loss": 3.0215,
|
| 156 |
+
"step": 210
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.004636264014161679,
|
| 160 |
+
"grad_norm": 3.368596315383911,
|
| 161 |
+
"learning_rate": 4.976924049565878e-05,
|
| 162 |
+
"loss": 3.096,
|
| 163 |
+
"step": 220
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.004847003287532664,
|
| 167 |
+
"grad_norm": 7.204585075378418,
|
| 168 |
+
"learning_rate": 4.975870353199023e-05,
|
| 169 |
+
"loss": 2.9822,
|
| 170 |
+
"step": 230
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.00505774256090365,
|
| 174 |
+
"grad_norm": 4.260004997253418,
|
| 175 |
+
"learning_rate": 4.9748166568321675e-05,
|
| 176 |
+
"loss": 3.0716,
|
| 177 |
+
"step": 240
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.005268481834274636,
|
| 181 |
+
"grad_norm": 4.109992504119873,
|
| 182 |
+
"learning_rate": 4.973762960465312e-05,
|
| 183 |
+
"loss": 2.9789,
|
| 184 |
+
"step": 250
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 0.005479221107645621,
|
| 188 |
+
"grad_norm": 10.826703071594238,
|
| 189 |
+
"learning_rate": 4.972709264098457e-05,
|
| 190 |
+
"loss": 2.9928,
|
| 191 |
+
"step": 260
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.005689960381016606,
|
| 195 |
+
"grad_norm": 4.802614212036133,
|
| 196 |
+
"learning_rate": 4.9716555677316027e-05,
|
| 197 |
+
"loss": 3.0421,
|
| 198 |
+
"step": 270
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 0.0059006996543875916,
|
| 202 |
+
"grad_norm": 4.688140392303467,
|
| 203 |
+
"learning_rate": 4.970601871364748e-05,
|
| 204 |
+
"loss": 2.8983,
|
| 205 |
+
"step": 280
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.006111438927758577,
|
| 209 |
+
"grad_norm": 4.237213134765625,
|
| 210 |
+
"learning_rate": 4.969548174997893e-05,
|
| 211 |
+
"loss": 3.0775,
|
| 212 |
+
"step": 290
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.006322178201129562,
|
| 216 |
+
"grad_norm": 6.8719000816345215,
|
| 217 |
+
"learning_rate": 4.968494478631038e-05,
|
| 218 |
+
"loss": 2.9356,
|
| 219 |
+
"step": 300
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.006532917474500548,
|
| 223 |
+
"grad_norm": 6.765329360961914,
|
| 224 |
+
"learning_rate": 4.9674407822641826e-05,
|
| 225 |
+
"loss": 3.0783,
|
| 226 |
+
"step": 310
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 0.006743656747871534,
|
| 230 |
+
"grad_norm": 4.296706199645996,
|
| 231 |
+
"learning_rate": 4.9664924555340135e-05,
|
| 232 |
+
"loss": 2.831,
|
| 233 |
+
"step": 320
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.006954396021242519,
|
| 237 |
+
"grad_norm": 5.505979537963867,
|
| 238 |
+
"learning_rate": 4.965438759167158e-05,
|
| 239 |
+
"loss": 3.0016,
|
| 240 |
+
"step": 330
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 0.007165135294613504,
|
| 244 |
+
"grad_norm": 15.19791316986084,
|
| 245 |
+
"learning_rate": 4.964385062800304e-05,
|
| 246 |
+
"loss": 2.9707,
|
| 247 |
+
"step": 340
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.0073758745679844894,
|
| 251 |
+
"grad_norm": 3.5731112957000732,
|
| 252 |
+
"learning_rate": 4.9633313664334487e-05,
|
| 253 |
+
"loss": 3.0052,
|
| 254 |
+
"step": 350
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.007586613841355475,
|
| 258 |
+
"grad_norm": 4.196691989898682,
|
| 259 |
+
"learning_rate": 4.962277670066594e-05,
|
| 260 |
+
"loss": 3.0018,
|
| 261 |
+
"step": 360
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.00779735311472646,
|
| 265 |
+
"grad_norm": 3.7209060192108154,
|
| 266 |
+
"learning_rate": 4.961223973699739e-05,
|
| 267 |
+
"loss": 2.9956,
|
| 268 |
+
"step": 370
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 0.008008092388097446,
|
| 272 |
+
"grad_norm": 4.67873477935791,
|
| 273 |
+
"learning_rate": 4.960170277332884e-05,
|
| 274 |
+
"loss": 3.0256,
|
| 275 |
+
"step": 380
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.008218831661468431,
|
| 279 |
+
"grad_norm": 4.471567630767822,
|
| 280 |
+
"learning_rate": 4.9591165809660286e-05,
|
| 281 |
+
"loss": 2.9656,
|
| 282 |
+
"step": 390
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.008429570934839417,
|
| 286 |
+
"grad_norm": 4.6043195724487305,
|
| 287 |
+
"learning_rate": 4.958062884599174e-05,
|
| 288 |
+
"loss": 2.9317,
|
| 289 |
+
"step": 400
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.008640310208210402,
|
| 293 |
+
"grad_norm": 5.873717308044434,
|
| 294 |
+
"learning_rate": 4.9570091882323196e-05,
|
| 295 |
+
"loss": 3.0048,
|
| 296 |
+
"step": 410
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.008851049481581387,
|
| 300 |
+
"grad_norm": 4.8152923583984375,
|
| 301 |
+
"learning_rate": 4.9559554918654645e-05,
|
| 302 |
+
"loss": 2.933,
|
| 303 |
+
"step": 420
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.009061788754952373,
|
| 307 |
+
"grad_norm": 3.463914632797241,
|
| 308 |
+
"learning_rate": 4.954901795498609e-05,
|
| 309 |
+
"loss": 2.9188,
|
| 310 |
+
"step": 430
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 0.009272528028323358,
|
| 314 |
+
"grad_norm": 5.305572986602783,
|
| 315 |
+
"learning_rate": 4.953848099131754e-05,
|
| 316 |
+
"loss": 3.1017,
|
| 317 |
+
"step": 440
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.009483267301694343,
|
| 321 |
+
"grad_norm": 3.027397394180298,
|
| 322 |
+
"learning_rate": 4.952794402764899e-05,
|
| 323 |
+
"loss": 2.9651,
|
| 324 |
+
"step": 450
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 0.009694006575065328,
|
| 328 |
+
"grad_norm": 3.619413375854492,
|
| 329 |
+
"learning_rate": 4.9517407063980444e-05,
|
| 330 |
+
"loss": 3.0447,
|
| 331 |
+
"step": 460
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.009904745848436314,
|
| 335 |
+
"grad_norm": 3.6718661785125732,
|
| 336 |
+
"learning_rate": 4.95068701003119e-05,
|
| 337 |
+
"loss": 2.9183,
|
| 338 |
+
"step": 470
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.0101154851218073,
|
| 342 |
+
"grad_norm": 3.527235984802246,
|
| 343 |
+
"learning_rate": 4.949633313664335e-05,
|
| 344 |
+
"loss": 2.9889,
|
| 345 |
+
"step": 480
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.010326224395178286,
|
| 349 |
+
"grad_norm": 3.9344675540924072,
|
| 350 |
+
"learning_rate": 4.9485796172974796e-05,
|
| 351 |
+
"loss": 2.9665,
|
| 352 |
+
"step": 490
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 0.010536963668549271,
|
| 356 |
+
"grad_norm": 6.459798336029053,
|
| 357 |
+
"learning_rate": 4.9475259209306244e-05,
|
| 358 |
+
"loss": 2.9079,
|
| 359 |
+
"step": 500
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.010747702941920257,
|
| 363 |
+
"grad_norm": 3.535411834716797,
|
| 364 |
+
"learning_rate": 4.94647222456377e-05,
|
| 365 |
+
"loss": 3.0064,
|
| 366 |
+
"step": 510
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 0.010958442215291242,
|
| 370 |
+
"grad_norm": 3.441504716873169,
|
| 371 |
+
"learning_rate": 4.9454185281969154e-05,
|
| 372 |
+
"loss": 2.8912,
|
| 373 |
+
"step": 520
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.011169181488662227,
|
| 377 |
+
"grad_norm": 6.144808292388916,
|
| 378 |
+
"learning_rate": 4.94436483183006e-05,
|
| 379 |
+
"loss": 2.9239,
|
| 380 |
+
"step": 530
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.011379920762033213,
|
| 384 |
+
"grad_norm": 3.7004387378692627,
|
| 385 |
+
"learning_rate": 4.943311135463205e-05,
|
| 386 |
+
"loss": 2.9526,
|
| 387 |
+
"step": 540
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.011590660035404198,
|
| 391 |
+
"grad_norm": 6.244487762451172,
|
| 392 |
+
"learning_rate": 4.94225743909635e-05,
|
| 393 |
+
"loss": 2.8755,
|
| 394 |
+
"step": 550
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 0.011801399308775183,
|
| 398 |
+
"grad_norm": 3.984456777572632,
|
| 399 |
+
"learning_rate": 4.9412037427294954e-05,
|
| 400 |
+
"loss": 2.835,
|
| 401 |
+
"step": 560
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.012012138582146168,
|
| 405 |
+
"grad_norm": 4.6722941398620605,
|
| 406 |
+
"learning_rate": 4.94015004636264e-05,
|
| 407 |
+
"loss": 2.9199,
|
| 408 |
+
"step": 570
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 0.012222877855517154,
|
| 412 |
+
"grad_norm": 4.149166584014893,
|
| 413 |
+
"learning_rate": 4.939096349995786e-05,
|
| 414 |
+
"loss": 3.0706,
|
| 415 |
+
"step": 580
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 0.012433617128888139,
|
| 419 |
+
"grad_norm": 3.566357374191284,
|
| 420 |
+
"learning_rate": 4.9380426536289305e-05,
|
| 421 |
+
"loss": 2.9011,
|
| 422 |
+
"step": 590
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.012644356402259124,
|
| 426 |
+
"grad_norm": 3.4722342491149902,
|
| 427 |
+
"learning_rate": 4.9369889572620754e-05,
|
| 428 |
+
"loss": 2.9345,
|
| 429 |
+
"step": 600
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.01285509567563011,
|
| 433 |
+
"grad_norm": 3.4608707427978516,
|
| 434 |
+
"learning_rate": 4.935935260895221e-05,
|
| 435 |
+
"loss": 2.8419,
|
| 436 |
+
"step": 610
|
| 437 |
+
},
|
| 438 |
+
{
|
| 439 |
+
"epoch": 0.013065834949001097,
|
| 440 |
+
"grad_norm": 3.4312050342559814,
|
| 441 |
+
"learning_rate": 4.934881564528366e-05,
|
| 442 |
+
"loss": 2.9087,
|
| 443 |
+
"step": 620
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.013276574222372082,
|
| 447 |
+
"grad_norm": 6.524892330169678,
|
| 448 |
+
"learning_rate": 4.9338278681615105e-05,
|
| 449 |
+
"loss": 2.8818,
|
| 450 |
+
"step": 630
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 0.013487313495743067,
|
| 454 |
+
"grad_norm": 5.508037567138672,
|
| 455 |
+
"learning_rate": 4.932774171794656e-05,
|
| 456 |
+
"loss": 2.7335,
|
| 457 |
+
"step": 640
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 0.013698052769114052,
|
| 461 |
+
"grad_norm": 4.094667911529541,
|
| 462 |
+
"learning_rate": 4.931720475427801e-05,
|
| 463 |
+
"loss": 2.9197,
|
| 464 |
+
"step": 650
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.013908792042485038,
|
| 468 |
+
"grad_norm": 3.6645474433898926,
|
| 469 |
+
"learning_rate": 4.9306667790609463e-05,
|
| 470 |
+
"loss": 2.9647,
|
| 471 |
+
"step": 660
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.014119531315856023,
|
| 475 |
+
"grad_norm": 3.711179256439209,
|
| 476 |
+
"learning_rate": 4.929613082694091e-05,
|
| 477 |
+
"loss": 2.9863,
|
| 478 |
+
"step": 670
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 0.014330270589227008,
|
| 482 |
+
"grad_norm": 3.904421329498291,
|
| 483 |
+
"learning_rate": 4.928559386327236e-05,
|
| 484 |
+
"loss": 2.8947,
|
| 485 |
+
"step": 680
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 0.014541009862597994,
|
| 489 |
+
"grad_norm": 3.3760502338409424,
|
| 490 |
+
"learning_rate": 4.927505689960381e-05,
|
| 491 |
+
"loss": 2.7375,
|
| 492 |
+
"step": 690
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 0.014751749135968979,
|
| 496 |
+
"grad_norm": 4.068516254425049,
|
| 497 |
+
"learning_rate": 4.926451993593526e-05,
|
| 498 |
+
"loss": 2.8502,
|
| 499 |
+
"step": 700
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 0.014962488409339964,
|
| 503 |
+
"grad_norm": 4.1656999588012695,
|
| 504 |
+
"learning_rate": 4.925398297226672e-05,
|
| 505 |
+
"loss": 2.8688,
|
| 506 |
+
"step": 710
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 0.01517322768271095,
|
| 510 |
+
"grad_norm": 3.5596961975097656,
|
| 511 |
+
"learning_rate": 4.9243446008598166e-05,
|
| 512 |
+
"loss": 2.9094,
|
| 513 |
+
"step": 720
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.015383966956081935,
|
| 517 |
+
"grad_norm": 4.3002753257751465,
|
| 518 |
+
"learning_rate": 4.9232909044929615e-05,
|
| 519 |
+
"loss": 2.8964,
|
| 520 |
+
"step": 730
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 0.01559470622945292,
|
| 524 |
+
"grad_norm": 5.381309509277344,
|
| 525 |
+
"learning_rate": 4.922237208126106e-05,
|
| 526 |
+
"loss": 2.8543,
|
| 527 |
+
"step": 740
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 0.015805445502823905,
|
| 531 |
+
"grad_norm": 5.13466739654541,
|
| 532 |
+
"learning_rate": 4.921183511759252e-05,
|
| 533 |
+
"loss": 2.8754,
|
| 534 |
+
"step": 750
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 0.016016184776194892,
|
| 538 |
+
"grad_norm": 4.221550464630127,
|
| 539 |
+
"learning_rate": 4.9201298153923966e-05,
|
| 540 |
+
"loss": 2.9192,
|
| 541 |
+
"step": 760
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 0.016226924049565876,
|
| 545 |
+
"grad_norm": 4.988082408905029,
|
| 546 |
+
"learning_rate": 4.919076119025542e-05,
|
| 547 |
+
"loss": 2.8325,
|
| 548 |
+
"step": 770
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 0.016437663322936863,
|
| 552 |
+
"grad_norm": 6.187538146972656,
|
| 553 |
+
"learning_rate": 4.918022422658687e-05,
|
| 554 |
+
"loss": 2.8819,
|
| 555 |
+
"step": 780
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.016648402596307846,
|
| 559 |
+
"grad_norm": 3.920051336288452,
|
| 560 |
+
"learning_rate": 4.916968726291832e-05,
|
| 561 |
+
"loss": 2.8123,
|
| 562 |
+
"step": 790
|
| 563 |
+
},
|
| 564 |
+
{
|
| 565 |
+
"epoch": 0.016859141869678834,
|
| 566 |
+
"grad_norm": 3.4514334201812744,
|
| 567 |
+
"learning_rate": 4.915915029924977e-05,
|
| 568 |
+
"loss": 2.8523,
|
| 569 |
+
"step": 800
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 0.017069881143049817,
|
| 573 |
+
"grad_norm": 5.007158279418945,
|
| 574 |
+
"learning_rate": 4.914861333558122e-05,
|
| 575 |
+
"loss": 2.8607,
|
| 576 |
+
"step": 810
|
| 577 |
+
},
|
| 578 |
+
{
|
| 579 |
+
"epoch": 0.017280620416420804,
|
| 580 |
+
"grad_norm": 4.900107383728027,
|
| 581 |
+
"learning_rate": 4.913807637191267e-05,
|
| 582 |
+
"loss": 2.7612,
|
| 583 |
+
"step": 820
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 0.01749135968979179,
|
| 587 |
+
"grad_norm": 7.45019006729126,
|
| 588 |
+
"learning_rate": 4.9127539408244124e-05,
|
| 589 |
+
"loss": 2.913,
|
| 590 |
+
"step": 830
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 0.017702098963162775,
|
| 594 |
+
"grad_norm": 5.6856770515441895,
|
| 595 |
+
"learning_rate": 4.911700244457557e-05,
|
| 596 |
+
"loss": 2.8877,
|
| 597 |
+
"step": 840
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 0.01791283823653376,
|
| 601 |
+
"grad_norm": 4.228877067565918,
|
| 602 |
+
"learning_rate": 4.910646548090703e-05,
|
| 603 |
+
"loss": 2.7294,
|
| 604 |
+
"step": 850
|
| 605 |
+
},
|
| 606 |
+
{
|
| 607 |
+
"epoch": 0.018123577509904745,
|
| 608 |
+
"grad_norm": 3.372474193572998,
|
| 609 |
+
"learning_rate": 4.9095928517238476e-05,
|
| 610 |
+
"loss": 2.7951,
|
| 611 |
+
"step": 860
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 0.018334316783275732,
|
| 615 |
+
"grad_norm": 3.3396689891815186,
|
| 616 |
+
"learning_rate": 4.9085391553569924e-05,
|
| 617 |
+
"loss": 2.8488,
|
| 618 |
+
"step": 870
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"epoch": 0.018545056056646716,
|
| 622 |
+
"grad_norm": 3.912947177886963,
|
| 623 |
+
"learning_rate": 4.907485458990138e-05,
|
| 624 |
+
"loss": 2.7816,
|
| 625 |
+
"step": 880
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 0.018755795330017703,
|
| 629 |
+
"grad_norm": 3.7946770191192627,
|
| 630 |
+
"learning_rate": 4.906431762623283e-05,
|
| 631 |
+
"loss": 2.8841,
|
| 632 |
+
"step": 890
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 0.018966534603388686,
|
| 636 |
+
"grad_norm": 4.324241638183594,
|
| 637 |
+
"learning_rate": 4.905378066256428e-05,
|
| 638 |
+
"loss": 2.8666,
|
| 639 |
+
"step": 900
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.019177273876759673,
|
| 643 |
+
"grad_norm": 3.172816038131714,
|
| 644 |
+
"learning_rate": 4.904324369889573e-05,
|
| 645 |
+
"loss": 2.7794,
|
| 646 |
+
"step": 910
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"epoch": 0.019388013150130657,
|
| 650 |
+
"grad_norm": 3.8265621662139893,
|
| 651 |
+
"learning_rate": 4.903270673522718e-05,
|
| 652 |
+
"loss": 2.8162,
|
| 653 |
+
"step": 920
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 0.019598752423501644,
|
| 657 |
+
"grad_norm": 3.4845385551452637,
|
| 658 |
+
"learning_rate": 4.902216977155863e-05,
|
| 659 |
+
"loss": 2.7385,
|
| 660 |
+
"step": 930
|
| 661 |
+
},
|
| 662 |
+
{
|
| 663 |
+
"epoch": 0.019809491696872628,
|
| 664 |
+
"grad_norm": 4.434839248657227,
|
| 665 |
+
"learning_rate": 4.901163280789008e-05,
|
| 666 |
+
"loss": 2.6926,
|
| 667 |
+
"step": 940
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 0.020020230970243615,
|
| 671 |
+
"grad_norm": 6.55767822265625,
|
| 672 |
+
"learning_rate": 4.900109584422154e-05,
|
| 673 |
+
"loss": 2.6728,
|
| 674 |
+
"step": 950
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 0.0202309702436146,
|
| 678 |
+
"grad_norm": 3.1376519203186035,
|
| 679 |
+
"learning_rate": 4.8990558880552985e-05,
|
| 680 |
+
"loss": 2.8854,
|
| 681 |
+
"step": 960
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 0.020441709516985585,
|
| 685 |
+
"grad_norm": 3.569209337234497,
|
| 686 |
+
"learning_rate": 4.8980021916884433e-05,
|
| 687 |
+
"loss": 2.7862,
|
| 688 |
+
"step": 970
|
| 689 |
+
},
|
| 690 |
+
{
|
| 691 |
+
"epoch": 0.020652448790356572,
|
| 692 |
+
"grad_norm": 16.01164436340332,
|
| 693 |
+
"learning_rate": 4.896948495321588e-05,
|
| 694 |
+
"loss": 2.8844,
|
| 695 |
+
"step": 980
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"epoch": 0.020863188063727556,
|
| 699 |
+
"grad_norm": 3.2311861515045166,
|
| 700 |
+
"learning_rate": 4.895894798954733e-05,
|
| 701 |
+
"loss": 2.8773,
|
| 702 |
+
"step": 990
|
| 703 |
+
},
|
| 704 |
+
{
|
| 705 |
+
"epoch": 0.021073927337098543,
|
| 706 |
+
"grad_norm": 4.047968864440918,
|
| 707 |
+
"learning_rate": 4.8948411025878785e-05,
|
| 708 |
+
"loss": 2.6955,
|
| 709 |
+
"step": 1000
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"epoch": 0.021284666610469526,
|
| 713 |
+
"grad_norm": 3.3213765621185303,
|
| 714 |
+
"learning_rate": 4.893787406221024e-05,
|
| 715 |
+
"loss": 2.8308,
|
| 716 |
+
"step": 1010
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 0.021495405883840513,
|
| 720 |
+
"grad_norm": 3.658327102661133,
|
| 721 |
+
"learning_rate": 4.892733709854169e-05,
|
| 722 |
+
"loss": 2.8409,
|
| 723 |
+
"step": 1020
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 0.021706145157211497,
|
| 727 |
+
"grad_norm": 3.6058104038238525,
|
| 728 |
+
"learning_rate": 4.8916800134873136e-05,
|
| 729 |
+
"loss": 2.8073,
|
| 730 |
+
"step": 1030
|
| 731 |
+
},
|
| 732 |
+
{
|
| 733 |
+
"epoch": 0.021916884430582484,
|
| 734 |
+
"grad_norm": 3.622807741165161,
|
| 735 |
+
"learning_rate": 4.8906263171204585e-05,
|
| 736 |
+
"loss": 2.7389,
|
| 737 |
+
"step": 1040
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 0.022127623703953467,
|
| 741 |
+
"grad_norm": 3.6087045669555664,
|
| 742 |
+
"learning_rate": 4.889572620753604e-05,
|
| 743 |
+
"loss": 2.7506,
|
| 744 |
+
"step": 1050
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 0.022338362977324455,
|
| 748 |
+
"grad_norm": 3.1047306060791016,
|
| 749 |
+
"learning_rate": 4.888518924386749e-05,
|
| 750 |
+
"loss": 2.77,
|
| 751 |
+
"step": 1060
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 0.022549102250695438,
|
| 755 |
+
"grad_norm": 3.4957761764526367,
|
| 756 |
+
"learning_rate": 4.887465228019894e-05,
|
| 757 |
+
"loss": 2.8526,
|
| 758 |
+
"step": 1070
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 0.022759841524066425,
|
| 762 |
+
"grad_norm": 5.486833095550537,
|
| 763 |
+
"learning_rate": 4.886411531653039e-05,
|
| 764 |
+
"loss": 2.8144,
|
| 765 |
+
"step": 1080
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 0.022970580797437412,
|
| 769 |
+
"grad_norm": 6.019243240356445,
|
| 770 |
+
"learning_rate": 4.885357835286184e-05,
|
| 771 |
+
"loss": 2.8095,
|
| 772 |
+
"step": 1090
|
| 773 |
+
},
|
| 774 |
+
{
|
| 775 |
+
"epoch": 0.023181320070808396,
|
| 776 |
+
"grad_norm": 5.631433486938477,
|
| 777 |
+
"learning_rate": 4.8843041389193294e-05,
|
| 778 |
+
"loss": 2.6904,
|
| 779 |
+
"step": 1100
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 0.023392059344179383,
|
| 783 |
+
"grad_norm": 3.162370204925537,
|
| 784 |
+
"learning_rate": 4.883250442552474e-05,
|
| 785 |
+
"loss": 2.6805,
|
| 786 |
+
"step": 1110
|
| 787 |
+
},
|
| 788 |
+
{
|
| 789 |
+
"epoch": 0.023602798617550366,
|
| 790 |
+
"grad_norm": 3.7855355739593506,
|
| 791 |
+
"learning_rate": 4.882196746185619e-05,
|
| 792 |
+
"loss": 2.7785,
|
| 793 |
+
"step": 1120
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 0.023813537890921353,
|
| 797 |
+
"grad_norm": 3.0686984062194824,
|
| 798 |
+
"learning_rate": 4.8811430498187646e-05,
|
| 799 |
+
"loss": 2.7163,
|
| 800 |
+
"step": 1130
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 0.024024277164292337,
|
| 804 |
+
"grad_norm": 4.290556907653809,
|
| 805 |
+
"learning_rate": 4.8800893534519094e-05,
|
| 806 |
+
"loss": 2.8549,
|
| 807 |
+
"step": 1140
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 0.024235016437663324,
|
| 811 |
+
"grad_norm": 4.224363327026367,
|
| 812 |
+
"learning_rate": 4.879035657085055e-05,
|
| 813 |
+
"loss": 2.7858,
|
| 814 |
+
"step": 1150
|
| 815 |
+
},
|
| 816 |
+
{
|
| 817 |
+
"epoch": 0.024445755711034307,
|
| 818 |
+
"grad_norm": 4.114129066467285,
|
| 819 |
+
"learning_rate": 4.8779819607182e-05,
|
| 820 |
+
"loss": 2.7969,
|
| 821 |
+
"step": 1160
|
| 822 |
+
},
|
| 823 |
+
{
|
| 824 |
+
"epoch": 0.024656494984405294,
|
| 825 |
+
"grad_norm": 4.066317081451416,
|
| 826 |
+
"learning_rate": 4.8769282643513446e-05,
|
| 827 |
+
"loss": 2.701,
|
| 828 |
+
"step": 1170
|
| 829 |
+
},
|
| 830 |
+
{
|
| 831 |
+
"epoch": 0.024867234257776278,
|
| 832 |
+
"grad_norm": 3.0110621452331543,
|
| 833 |
+
"learning_rate": 4.8758745679844894e-05,
|
| 834 |
+
"loss": 2.9004,
|
| 835 |
+
"step": 1180
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"epoch": 0.025077973531147265,
|
| 839 |
+
"grad_norm": 4.144210338592529,
|
| 840 |
+
"learning_rate": 4.874820871617635e-05,
|
| 841 |
+
"loss": 2.8651,
|
| 842 |
+
"step": 1190
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 0.02528871280451825,
|
| 846 |
+
"grad_norm": 3.226668119430542,
|
| 847 |
+
"learning_rate": 4.8737671752507804e-05,
|
| 848 |
+
"loss": 2.8332,
|
| 849 |
+
"step": 1200
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 0.025499452077889236,
|
| 853 |
+
"grad_norm": 3.7419745922088623,
|
| 854 |
+
"learning_rate": 4.872713478883925e-05,
|
| 855 |
+
"loss": 2.8502,
|
| 856 |
+
"step": 1210
|
| 857 |
+
},
|
| 858 |
+
{
|
| 859 |
+
"epoch": 0.02571019135126022,
|
| 860 |
+
"grad_norm": 6.740761756896973,
|
| 861 |
+
"learning_rate": 4.87165978251707e-05,
|
| 862 |
+
"loss": 2.7892,
|
| 863 |
+
"step": 1220
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"epoch": 0.025920930624631206,
|
| 867 |
+
"grad_norm": 4.075252056121826,
|
| 868 |
+
"learning_rate": 4.870606086150215e-05,
|
| 869 |
+
"loss": 2.7037,
|
| 870 |
+
"step": 1230
|
| 871 |
+
},
|
| 872 |
+
{
|
| 873 |
+
"epoch": 0.026131669898002193,
|
| 874 |
+
"grad_norm": 3.7171854972839355,
|
| 875 |
+
"learning_rate": 4.8695523897833604e-05,
|
| 876 |
+
"loss": 2.7885,
|
| 877 |
+
"step": 1240
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"epoch": 0.026342409171373177,
|
| 881 |
+
"grad_norm": 3.2372097969055176,
|
| 882 |
+
"learning_rate": 4.868498693416506e-05,
|
| 883 |
+
"loss": 2.7827,
|
| 884 |
+
"step": 1250
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 0.026553148444744164,
|
| 888 |
+
"grad_norm": 5.390050888061523,
|
| 889 |
+
"learning_rate": 4.867444997049651e-05,
|
| 890 |
+
"loss": 2.7929,
|
| 891 |
+
"step": 1260
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 0.026763887718115147,
|
| 895 |
+
"grad_norm": 6.445965766906738,
|
| 896 |
+
"learning_rate": 4.8663913006827955e-05,
|
| 897 |
+
"loss": 2.7683,
|
| 898 |
+
"step": 1270
|
| 899 |
+
},
|
| 900 |
+
{
|
| 901 |
+
"epoch": 0.026974626991486134,
|
| 902 |
+
"grad_norm": 4.038309574127197,
|
| 903 |
+
"learning_rate": 4.8653376043159403e-05,
|
| 904 |
+
"loss": 2.7644,
|
| 905 |
+
"step": 1280
|
| 906 |
+
},
|
| 907 |
+
{
|
| 908 |
+
"epoch": 0.027185366264857118,
|
| 909 |
+
"grad_norm": 3.422018527984619,
|
| 910 |
+
"learning_rate": 4.864283907949085e-05,
|
| 911 |
+
"loss": 2.759,
|
| 912 |
+
"step": 1290
|
| 913 |
+
},
|
| 914 |
+
{
|
| 915 |
+
"epoch": 0.027396105538228105,
|
| 916 |
+
"grad_norm": 3.4551455974578857,
|
| 917 |
+
"learning_rate": 4.863230211582231e-05,
|
| 918 |
+
"loss": 2.6919,
|
| 919 |
+
"step": 1300
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"epoch": 0.02760684481159909,
|
| 923 |
+
"grad_norm": 4.99464225769043,
|
| 924 |
+
"learning_rate": 4.862176515215376e-05,
|
| 925 |
+
"loss": 2.8362,
|
| 926 |
+
"step": 1310
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 0.027817584084970075,
|
| 930 |
+
"grad_norm": 3.759676933288574,
|
| 931 |
+
"learning_rate": 4.861122818848521e-05,
|
| 932 |
+
"loss": 2.7712,
|
| 933 |
+
"step": 1320
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 0.02802832335834106,
|
| 937 |
+
"grad_norm": 4.082064151763916,
|
| 938 |
+
"learning_rate": 4.860069122481666e-05,
|
| 939 |
+
"loss": 2.8224,
|
| 940 |
+
"step": 1330
|
| 941 |
+
},
|
| 942 |
+
{
|
| 943 |
+
"epoch": 0.028239062631712046,
|
| 944 |
+
"grad_norm": 3.157449960708618,
|
| 945 |
+
"learning_rate": 4.8590154261148106e-05,
|
| 946 |
+
"loss": 2.7863,
|
| 947 |
+
"step": 1340
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 0.02844980190508303,
|
| 951 |
+
"grad_norm": 5.148242950439453,
|
| 952 |
+
"learning_rate": 4.857961729747956e-05,
|
| 953 |
+
"loss": 2.7383,
|
| 954 |
+
"step": 1350
|
| 955 |
+
},
|
| 956 |
+
{
|
| 957 |
+
"epoch": 0.028660541178454017,
|
| 958 |
+
"grad_norm": 3.794858694076538,
|
| 959 |
+
"learning_rate": 4.856908033381101e-05,
|
| 960 |
+
"loss": 2.7118,
|
| 961 |
+
"step": 1360
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"epoch": 0.028871280451825004,
|
| 965 |
+
"grad_norm": 3.0035576820373535,
|
| 966 |
+
"learning_rate": 4.8558543370142465e-05,
|
| 967 |
+
"loss": 2.7787,
|
| 968 |
+
"step": 1370
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 0.029082019725195987,
|
| 972 |
+
"grad_norm": 4.196252822875977,
|
| 973 |
+
"learning_rate": 4.854800640647391e-05,
|
| 974 |
+
"loss": 2.7501,
|
| 975 |
+
"step": 1380
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 0.029292758998566974,
|
| 979 |
+
"grad_norm": 7.164214611053467,
|
| 980 |
+
"learning_rate": 4.853746944280536e-05,
|
| 981 |
+
"loss": 2.7885,
|
| 982 |
+
"step": 1390
|
| 983 |
+
},
|
| 984 |
+
{
|
| 985 |
+
"epoch": 0.029503498271937958,
|
| 986 |
+
"grad_norm": 3.252230167388916,
|
| 987 |
+
"learning_rate": 4.8526932479136816e-05,
|
| 988 |
+
"loss": 2.7875,
|
| 989 |
+
"step": 1400
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"epoch": 0.029714237545308945,
|
| 993 |
+
"grad_norm": 7.037599563598633,
|
| 994 |
+
"learning_rate": 4.8516395515468265e-05,
|
| 995 |
+
"loss": 2.7787,
|
| 996 |
+
"step": 1410
|
| 997 |
+
},
|
| 998 |
+
{
|
| 999 |
+
"epoch": 0.02992497681867993,
|
| 1000 |
+
"grad_norm": 3.5410661697387695,
|
| 1001 |
+
"learning_rate": 4.850585855179971e-05,
|
| 1002 |
+
"loss": 2.8644,
|
| 1003 |
+
"step": 1420
|
| 1004 |
+
},
|
| 1005 |
+
{
|
| 1006 |
+
"epoch": 0.030135716092050915,
|
| 1007 |
+
"grad_norm": 3.7410356998443604,
|
| 1008 |
+
"learning_rate": 4.849532158813117e-05,
|
| 1009 |
+
"loss": 2.8146,
|
| 1010 |
+
"step": 1430
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 0.0303464553654219,
|
| 1014 |
+
"grad_norm": 7.049551963806152,
|
| 1015 |
+
"learning_rate": 4.8484784624462616e-05,
|
| 1016 |
+
"loss": 2.8427,
|
| 1017 |
+
"step": 1440
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 0.030557194638792886,
|
| 1021 |
+
"grad_norm": 3.5179500579833984,
|
| 1022 |
+
"learning_rate": 4.847424766079407e-05,
|
| 1023 |
+
"loss": 2.7396,
|
| 1024 |
+
"step": 1450
|
| 1025 |
+
},
|
| 1026 |
+
{
|
| 1027 |
+
"epoch": 0.03076793391216387,
|
| 1028 |
+
"grad_norm": 3.117222785949707,
|
| 1029 |
+
"learning_rate": 4.846371069712552e-05,
|
| 1030 |
+
"loss": 2.6701,
|
| 1031 |
+
"step": 1460
|
| 1032 |
+
},
|
| 1033 |
+
{
|
| 1034 |
+
"epoch": 0.030978673185534857,
|
| 1035 |
+
"grad_norm": 2.9434943199157715,
|
| 1036 |
+
"learning_rate": 4.845317373345697e-05,
|
| 1037 |
+
"loss": 2.7827,
|
| 1038 |
+
"step": 1470
|
| 1039 |
+
},
|
| 1040 |
+
{
|
| 1041 |
+
"epoch": 0.03118941245890584,
|
| 1042 |
+
"grad_norm": 7.40554666519165,
|
| 1043 |
+
"learning_rate": 4.8442636769788416e-05,
|
| 1044 |
+
"loss": 2.6663,
|
| 1045 |
+
"step": 1480
|
| 1046 |
+
},
|
| 1047 |
+
{
|
| 1048 |
+
"epoch": 0.031400151732276824,
|
| 1049 |
+
"grad_norm": 3.6940155029296875,
|
| 1050 |
+
"learning_rate": 4.843209980611987e-05,
|
| 1051 |
+
"loss": 2.7505,
|
| 1052 |
+
"step": 1490
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 0.03161089100564781,
|
| 1056 |
+
"grad_norm": 3.142557382583618,
|
| 1057 |
+
"learning_rate": 4.8421562842451326e-05,
|
| 1058 |
+
"loss": 2.7381,
|
| 1059 |
+
"step": 1500
|
| 1060 |
+
}
|
| 1061 |
+
],
|
| 1062 |
+
"logging_steps": 10,
|
| 1063 |
+
"max_steps": 47452,
|
| 1064 |
+
"num_input_tokens_seen": 0,
|
| 1065 |
+
"num_train_epochs": 1,
|
| 1066 |
+
"save_steps": 100,
|
| 1067 |
+
"stateful_callbacks": {
|
| 1068 |
+
"TrainerControl": {
|
| 1069 |
+
"args": {
|
| 1070 |
+
"should_epoch_stop": false,
|
| 1071 |
+
"should_evaluate": false,
|
| 1072 |
+
"should_log": false,
|
| 1073 |
+
"should_save": true,
|
| 1074 |
+
"should_training_stop": false
|
| 1075 |
+
},
|
| 1076 |
+
"attributes": {}
|
| 1077 |
+
}
|
| 1078 |
+
},
|
| 1079 |
+
"total_flos": 3376102996231680.0,
|
| 1080 |
+
"train_batch_size": 1,
|
| 1081 |
+
"trial_name": null,
|
| 1082 |
+
"trial_params": null
|
| 1083 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ed5bb1ef57b2ac804400da520cfcbeda430ae350f09924695fa4f0acd5855679
|
| 3 |
+
size 5304
|
vocab.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|