Commit
•
2a28cce
1
Parent(s):
eda4e07
Upload folder using huggingface_hub
Browse files- README.md +9 -0
- adapter_config.json +26 -0
- adapter_model.bin +3 -0
- config.json +45 -0
- non_lora_trainables.bin +3 -0
- trainer_state.json +1231 -0
README.md
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
library_name: peft
|
| 3 |
+
---
|
| 4 |
+
## Training procedure
|
| 5 |
+
|
| 6 |
+
### Framework versions
|
| 7 |
+
|
| 8 |
+
|
| 9 |
+
- PEFT 0.4.0
|
adapter_config.json
ADDED
|
@@ -0,0 +1,26 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"auto_mapping": null,
|
| 3 |
+
"base_model_name_or_path": "liuhaotian/llava-v1.5-13b",
|
| 4 |
+
"bias": "none",
|
| 5 |
+
"fan_in_fan_out": false,
|
| 6 |
+
"inference_mode": true,
|
| 7 |
+
"init_lora_weights": true,
|
| 8 |
+
"layers_pattern": null,
|
| 9 |
+
"layers_to_transform": null,
|
| 10 |
+
"lora_alpha": 256,
|
| 11 |
+
"lora_dropout": 0.05,
|
| 12 |
+
"modules_to_save": null,
|
| 13 |
+
"peft_type": "LORA",
|
| 14 |
+
"r": 128,
|
| 15 |
+
"revision": null,
|
| 16 |
+
"target_modules": [
|
| 17 |
+
"up_proj",
|
| 18 |
+
"q_proj",
|
| 19 |
+
"gate_proj",
|
| 20 |
+
"v_proj",
|
| 21 |
+
"down_proj",
|
| 22 |
+
"o_proj",
|
| 23 |
+
"k_proj"
|
| 24 |
+
],
|
| 25 |
+
"task_type": "CAUSAL_LM"
|
| 26 |
+
}
|
adapter_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:97c3df2774501a76b4c9a60033ebe3e1f04ee841f0a556dad6e2521629a4c997
|
| 3 |
+
size 1001584333
|
config.json
ADDED
|
@@ -0,0 +1,45 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "liuhaotian/llava-v1.5-13b",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"LlavaLlamaForCausalLM"
|
| 5 |
+
],
|
| 6 |
+
"bos_token_id": 1,
|
| 7 |
+
"eos_token_id": 2,
|
| 8 |
+
"freeze_mm_mlp_adapter": false,
|
| 9 |
+
"freeze_mm_vision_resampler": false,
|
| 10 |
+
"hidden_act": "silu",
|
| 11 |
+
"hidden_size": 5120,
|
| 12 |
+
"image_aspect_ratio": "pad",
|
| 13 |
+
"initializer_range": 0.02,
|
| 14 |
+
"intermediate_size": 13824,
|
| 15 |
+
"max_length": 4096,
|
| 16 |
+
"max_position_embeddings": 4096,
|
| 17 |
+
"mm_hidden_size": 1024,
|
| 18 |
+
"mm_projector_lr": 2e-05,
|
| 19 |
+
"mm_projector_type": "mlp2x_gelu",
|
| 20 |
+
"mm_resampler_type": null,
|
| 21 |
+
"mm_use_im_patch_token": false,
|
| 22 |
+
"mm_use_im_start_end": false,
|
| 23 |
+
"mm_vision_select_feature": "patch",
|
| 24 |
+
"mm_vision_select_layer": -2,
|
| 25 |
+
"mm_vision_tower": "openai/clip-vit-large-patch14-336",
|
| 26 |
+
"model_type": "llava",
|
| 27 |
+
"num_attention_heads": 40,
|
| 28 |
+
"num_hidden_layers": 40,
|
| 29 |
+
"num_key_value_heads": 40,
|
| 30 |
+
"pad_token_id": 0,
|
| 31 |
+
"pretraining_tp": 1,
|
| 32 |
+
"rms_norm_eps": 1e-05,
|
| 33 |
+
"rope_scaling": null,
|
| 34 |
+
"tie_word_embeddings": false,
|
| 35 |
+
"tokenizer_model_max_length": 2048,
|
| 36 |
+
"tokenizer_padding_side": "right",
|
| 37 |
+
"torch_dtype": "float16",
|
| 38 |
+
"transformers_version": "4.31.0",
|
| 39 |
+
"tune_mm_mlp_adapter": false,
|
| 40 |
+
"tune_mm_vision_resampler": false,
|
| 41 |
+
"unfreeze_mm_vision_tower": false,
|
| 42 |
+
"use_cache": true,
|
| 43 |
+
"use_mm_proj": true,
|
| 44 |
+
"vocab_size": 32000
|
| 45 |
+
}
|
non_lora_trainables.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d15449718903780fbeed63f4e60a81336bf65886be39a5636fe541cf8dec0685
|
| 3 |
+
size 62936807
|
trainer_state.json
ADDED
|
@@ -0,0 +1,1231 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 3.0,
|
| 5 |
+
"global_step": 201,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.01,
|
| 12 |
+
"learning_rate": 5.714285714285714e-05,
|
| 13 |
+
"loss": 1.5581,
|
| 14 |
+
"step": 1
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.03,
|
| 18 |
+
"learning_rate": 0.00011428571428571428,
|
| 19 |
+
"loss": 1.5618,
|
| 20 |
+
"step": 2
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.04,
|
| 24 |
+
"learning_rate": 0.00017142857142857143,
|
| 25 |
+
"loss": 1.4337,
|
| 26 |
+
"step": 3
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.06,
|
| 30 |
+
"learning_rate": 0.00022857142857142857,
|
| 31 |
+
"loss": 1.2552,
|
| 32 |
+
"step": 4
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.07,
|
| 36 |
+
"learning_rate": 0.00028571428571428574,
|
| 37 |
+
"loss": 1.2225,
|
| 38 |
+
"step": 5
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.09,
|
| 42 |
+
"learning_rate": 0.00034285714285714285,
|
| 43 |
+
"loss": 1.1791,
|
| 44 |
+
"step": 6
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.1,
|
| 48 |
+
"learning_rate": 0.0004,
|
| 49 |
+
"loss": 1.1597,
|
| 50 |
+
"step": 7
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.12,
|
| 54 |
+
"learning_rate": 0.00039997377673312647,
|
| 55 |
+
"loss": 1.1362,
|
| 56 |
+
"step": 8
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.13,
|
| 60 |
+
"learning_rate": 0.00039989511380910303,
|
| 61 |
+
"loss": 1.1049,
|
| 62 |
+
"step": 9
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.15,
|
| 66 |
+
"learning_rate": 0.0003997640318559182,
|
| 67 |
+
"loss": 1.0949,
|
| 68 |
+
"step": 10
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.16,
|
| 72 |
+
"learning_rate": 0.00039958056524754237,
|
| 73 |
+
"loss": 1.093,
|
| 74 |
+
"step": 11
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.18,
|
| 78 |
+
"learning_rate": 0.0003993447620949139,
|
| 79 |
+
"loss": 1.0893,
|
| 80 |
+
"step": 12
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.19,
|
| 84 |
+
"learning_rate": 0.0003990566842333228,
|
| 85 |
+
"loss": 1.1027,
|
| 86 |
+
"step": 13
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.21,
|
| 90 |
+
"learning_rate": 0.00039871640720619554,
|
| 91 |
+
"loss": 1.0704,
|
| 92 |
+
"step": 14
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.22,
|
| 96 |
+
"learning_rate": 0.0003983240202452851,
|
| 97 |
+
"loss": 1.0495,
|
| 98 |
+
"step": 15
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.24,
|
| 102 |
+
"learning_rate": 0.00039787962624727123,
|
| 103 |
+
"loss": 1.0833,
|
| 104 |
+
"step": 16
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 0.25,
|
| 108 |
+
"learning_rate": 0.00039738334174677813,
|
| 109 |
+
"loss": 1.0664,
|
| 110 |
+
"step": 17
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.27,
|
| 114 |
+
"learning_rate": 0.0003968352968858149,
|
| 115 |
+
"loss": 1.0566,
|
| 116 |
+
"step": 18
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 0.28,
|
| 120 |
+
"learning_rate": 0.0003962356353796478,
|
| 121 |
+
"loss": 1.0466,
|
| 122 |
+
"step": 19
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.3,
|
| 126 |
+
"learning_rate": 0.00039558451447911414,
|
| 127 |
+
"loss": 1.0814,
|
| 128 |
+
"step": 20
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.31,
|
| 132 |
+
"learning_rate": 0.00039488210492938527,
|
| 133 |
+
"loss": 1.0296,
|
| 134 |
+
"step": 21
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.33,
|
| 138 |
+
"learning_rate": 0.00039412859092519185,
|
| 139 |
+
"loss": 1.0381,
|
| 140 |
+
"step": 22
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 0.34,
|
| 144 |
+
"learning_rate": 0.00039332417006252224,
|
| 145 |
+
"loss": 1.0213,
|
| 146 |
+
"step": 23
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 0.36,
|
| 150 |
+
"learning_rate": 0.0003924690532868061,
|
| 151 |
+
"loss": 1.0226,
|
| 152 |
+
"step": 24
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 0.37,
|
| 156 |
+
"learning_rate": 0.00039156346483759744,
|
| 157 |
+
"loss": 1.061,
|
| 158 |
+
"step": 25
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 0.39,
|
| 162 |
+
"learning_rate": 0.0003906076421897722,
|
| 163 |
+
"loss": 1.059,
|
| 164 |
+
"step": 26
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.4,
|
| 168 |
+
"learning_rate": 0.0003896018359912541,
|
| 169 |
+
"loss": 1.0453,
|
| 170 |
+
"step": 27
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.42,
|
| 174 |
+
"learning_rate": 0.0003885463099972869,
|
| 175 |
+
"loss": 1.0195,
|
| 176 |
+
"step": 28
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 0.43,
|
| 180 |
+
"learning_rate": 0.00038744134100126876,
|
| 181 |
+
"loss": 1.0416,
|
| 182 |
+
"step": 29
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 0.45,
|
| 186 |
+
"learning_rate": 0.0003862872187621685,
|
| 187 |
+
"loss": 1.0315,
|
| 188 |
+
"step": 30
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 0.46,
|
| 192 |
+
"learning_rate": 0.0003850842459285409,
|
| 193 |
+
"loss": 1.0249,
|
| 194 |
+
"step": 31
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 0.48,
|
| 198 |
+
"learning_rate": 0.00038383273795916244,
|
| 199 |
+
"loss": 1.0359,
|
| 200 |
+
"step": 32
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 0.49,
|
| 204 |
+
"learning_rate": 0.00038253302304030806,
|
| 205 |
+
"loss": 1.0188,
|
| 206 |
+
"step": 33
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.51,
|
| 210 |
+
"learning_rate": 0.0003811854419996894,
|
| 211 |
+
"loss": 1.0426,
|
| 212 |
+
"step": 34
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.52,
|
| 216 |
+
"learning_rate": 0.0003797903482170791,
|
| 217 |
+
"loss": 1.0243,
|
| 218 |
+
"step": 35
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 0.54,
|
| 222 |
+
"learning_rate": 0.0003783481075316429,
|
| 223 |
+
"loss": 1.0176,
|
| 224 |
+
"step": 36
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 0.55,
|
| 228 |
+
"learning_rate": 0.00037685909814600473,
|
| 229 |
+
"loss": 1.0412,
|
| 230 |
+
"step": 37
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 0.57,
|
| 234 |
+
"learning_rate": 0.0003753237105270696,
|
| 235 |
+
"loss": 1.0295,
|
| 236 |
+
"step": 38
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 0.58,
|
| 240 |
+
"learning_rate": 0.00037374234730363023,
|
| 241 |
+
"loss": 1.0164,
|
| 242 |
+
"step": 39
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 0.6,
|
| 246 |
+
"learning_rate": 0.00037211542316078507,
|
| 247 |
+
"loss": 1.0196,
|
| 248 |
+
"step": 40
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 0.61,
|
| 252 |
+
"learning_rate": 0.0003704433647311939,
|
| 253 |
+
"loss": 1.0196,
|
| 254 |
+
"step": 41
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.63,
|
| 258 |
+
"learning_rate": 0.00036872661048320093,
|
| 259 |
+
"loss": 1.0177,
|
| 260 |
+
"step": 42
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 0.64,
|
| 264 |
+
"learning_rate": 0.00036696561060585426,
|
| 265 |
+
"loss": 1.0258,
|
| 266 |
+
"step": 43
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 0.66,
|
| 270 |
+
"learning_rate": 0.0003651608268908513,
|
| 271 |
+
"loss": 1.0212,
|
| 272 |
+
"step": 44
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 0.67,
|
| 276 |
+
"learning_rate": 0.0003633127326114422,
|
| 277 |
+
"loss": 1.0448,
|
| 278 |
+
"step": 45
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 0.69,
|
| 282 |
+
"learning_rate": 0.00036142181239832185,
|
| 283 |
+
"loss": 0.992,
|
| 284 |
+
"step": 46
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 0.7,
|
| 288 |
+
"learning_rate": 0.00035948856211254416,
|
| 289 |
+
"loss": 0.9978,
|
| 290 |
+
"step": 47
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 0.72,
|
| 294 |
+
"learning_rate": 0.00035751348871549093,
|
| 295 |
+
"loss": 1.0349,
|
| 296 |
+
"step": 48
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.73,
|
| 300 |
+
"learning_rate": 0.0003554971101359299,
|
| 301 |
+
"loss": 1.0298,
|
| 302 |
+
"step": 49
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 0.75,
|
| 306 |
+
"learning_rate": 0.00035343995513419725,
|
| 307 |
+
"loss": 1.0317,
|
| 308 |
+
"step": 50
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 0.76,
|
| 312 |
+
"learning_rate": 0.00035134256316353906,
|
| 313 |
+
"loss": 1.0056,
|
| 314 |
+
"step": 51
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 0.78,
|
| 318 |
+
"learning_rate": 0.0003492054842286492,
|
| 319 |
+
"loss": 1.0154,
|
| 320 |
+
"step": 52
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 0.79,
|
| 324 |
+
"learning_rate": 0.0003470292787414401,
|
| 325 |
+
"loss": 1.0194,
|
| 326 |
+
"step": 53
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 0.81,
|
| 330 |
+
"learning_rate": 0.00034481451737408437,
|
| 331 |
+
"loss": 0.9889,
|
| 332 |
+
"step": 54
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 0.82,
|
| 336 |
+
"learning_rate": 0.0003425617809093659,
|
| 337 |
+
"loss": 1.0256,
|
| 338 |
+
"step": 55
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.84,
|
| 342 |
+
"learning_rate": 0.00034027166008837985,
|
| 343 |
+
"loss": 0.9959,
|
| 344 |
+
"step": 56
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 0.85,
|
| 348 |
+
"learning_rate": 0.0003379447554556209,
|
| 349 |
+
"loss": 1.026,
|
| 350 |
+
"step": 57
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 0.87,
|
| 354 |
+
"learning_rate": 0.00033558167720150066,
|
| 355 |
+
"loss": 1.0113,
|
| 356 |
+
"step": 58
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 0.88,
|
| 360 |
+
"learning_rate": 0.00033318304500233625,
|
| 361 |
+
"loss": 1.0133,
|
| 362 |
+
"step": 59
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 0.9,
|
| 366 |
+
"learning_rate": 0.00033074948785785055,
|
| 367 |
+
"loss": 1.0133,
|
| 368 |
+
"step": 60
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 0.91,
|
| 372 |
+
"learning_rate": 0.000328281643926228,
|
| 373 |
+
"loss": 1.0156,
|
| 374 |
+
"step": 61
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 0.93,
|
| 378 |
+
"learning_rate": 0.0003257801603567689,
|
| 379 |
+
"loss": 1.0207,
|
| 380 |
+
"step": 62
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.94,
|
| 384 |
+
"learning_rate": 0.0003232456931201855,
|
| 385 |
+
"loss": 0.9841,
|
| 386 |
+
"step": 63
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 0.96,
|
| 390 |
+
"learning_rate": 0.00032067890683658496,
|
| 391 |
+
"loss": 1.0047,
|
| 392 |
+
"step": 64
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 0.97,
|
| 396 |
+
"learning_rate": 0.00031808047460118457,
|
| 397 |
+
"loss": 1.0015,
|
| 398 |
+
"step": 65
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 0.99,
|
| 402 |
+
"learning_rate": 0.00031545107780780394,
|
| 403 |
+
"loss": 0.9785,
|
| 404 |
+
"step": 66
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 1.0,
|
| 408 |
+
"learning_rate": 0.0003127914059701813,
|
| 409 |
+
"loss": 1.0009,
|
| 410 |
+
"step": 67
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 1.01,
|
| 414 |
+
"learning_rate": 0.0003101021565411607,
|
| 415 |
+
"loss": 0.8419,
|
| 416 |
+
"step": 68
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 1.03,
|
| 420 |
+
"learning_rate": 0.0003073840347297968,
|
| 421 |
+
"loss": 0.8403,
|
| 422 |
+
"step": 69
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 1.04,
|
| 426 |
+
"learning_rate": 0.000304637753316426,
|
| 427 |
+
"loss": 0.8149,
|
| 428 |
+
"step": 70
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 1.06,
|
| 432 |
+
"learning_rate": 0.00030186403246575263,
|
| 433 |
+
"loss": 0.8348,
|
| 434 |
+
"step": 71
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 1.07,
|
| 438 |
+
"learning_rate": 0.00029906359953799756,
|
| 439 |
+
"loss": 0.8394,
|
| 440 |
+
"step": 72
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 1.09,
|
| 444 |
+
"learning_rate": 0.00029623718889816103,
|
| 445 |
+
"loss": 0.8091,
|
| 446 |
+
"step": 73
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 1.1,
|
| 450 |
+
"learning_rate": 0.0002933855417234481,
|
| 451 |
+
"loss": 0.8093,
|
| 452 |
+
"step": 74
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 1.12,
|
| 456 |
+
"learning_rate": 0.00029050940580890785,
|
| 457 |
+
"loss": 0.7727,
|
| 458 |
+
"step": 75
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 1.13,
|
| 462 |
+
"learning_rate": 0.0002876095353713365,
|
| 463 |
+
"loss": 0.8193,
|
| 464 |
+
"step": 76
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 1.15,
|
| 468 |
+
"learning_rate": 0.0002846866908514981,
|
| 469 |
+
"loss": 0.8253,
|
| 470 |
+
"step": 77
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 1.16,
|
| 474 |
+
"learning_rate": 0.0002817416387147113,
|
| 475 |
+
"loss": 0.8263,
|
| 476 |
+
"step": 78
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 1.18,
|
| 480 |
+
"learning_rate": 0.00027877515124985746,
|
| 481 |
+
"loss": 0.8086,
|
| 482 |
+
"step": 79
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 1.19,
|
| 486 |
+
"learning_rate": 0.00027578800636686137,
|
| 487 |
+
"loss": 0.8282,
|
| 488 |
+
"step": 80
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 1.21,
|
| 492 |
+
"learning_rate": 0.0002727809873926975,
|
| 493 |
+
"loss": 0.7946,
|
| 494 |
+
"step": 81
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 1.22,
|
| 498 |
+
"learning_rate": 0.00026975488286597643,
|
| 499 |
+
"loss": 0.7947,
|
| 500 |
+
"step": 82
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 1.24,
|
| 504 |
+
"learning_rate": 0.00026671048633016415,
|
| 505 |
+
"loss": 0.8287,
|
| 506 |
+
"step": 83
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 1.25,
|
| 510 |
+
"learning_rate": 0.0002636485961254888,
|
| 511 |
+
"loss": 0.8201,
|
| 512 |
+
"step": 84
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 1.27,
|
| 516 |
+
"learning_rate": 0.0002605700151795901,
|
| 517 |
+
"loss": 0.7945,
|
| 518 |
+
"step": 85
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 1.28,
|
| 522 |
+
"learning_rate": 0.00025747555079696566,
|
| 523 |
+
"loss": 0.7827,
|
| 524 |
+
"step": 86
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 1.3,
|
| 528 |
+
"learning_rate": 0.0002543660144472686,
|
| 529 |
+
"loss": 0.8353,
|
| 530 |
+
"step": 87
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 1.31,
|
| 534 |
+
"learning_rate": 0.00025124222155251444,
|
| 535 |
+
"loss": 0.8163,
|
| 536 |
+
"step": 88
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 1.33,
|
| 540 |
+
"learning_rate": 0.00024810499127325073,
|
| 541 |
+
"loss": 0.7853,
|
| 542 |
+
"step": 89
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 1.34,
|
| 546 |
+
"learning_rate": 0.00024495514629374593,
|
| 547 |
+
"loss": 0.7858,
|
| 548 |
+
"step": 90
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 1.36,
|
| 552 |
+
"learning_rate": 0.00024179351260625507,
|
| 553 |
+
"loss": 0.7868,
|
| 554 |
+
"step": 91
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 1.37,
|
| 558 |
+
"learning_rate": 0.00023862091929441762,
|
| 559 |
+
"loss": 0.8161,
|
| 560 |
+
"step": 92
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 1.39,
|
| 564 |
+
"learning_rate": 0.0002354381983158446,
|
| 565 |
+
"loss": 0.7799,
|
| 566 |
+
"step": 93
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 1.4,
|
| 570 |
+
"learning_rate": 0.00023224618428395198,
|
| 571 |
+
"loss": 0.8114,
|
| 572 |
+
"step": 94
|
| 573 |
+
},
|
| 574 |
+
{
|
| 575 |
+
"epoch": 1.42,
|
| 576 |
+
"learning_rate": 0.00022904571424909808,
|
| 577 |
+
"loss": 0.8078,
|
| 578 |
+
"step": 95
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 1.43,
|
| 582 |
+
"learning_rate": 0.0002258376274790813,
|
| 583 |
+
"loss": 0.8077,
|
| 584 |
+
"step": 96
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 1.45,
|
| 588 |
+
"learning_rate": 0.0002226227652390569,
|
| 589 |
+
"loss": 0.7888,
|
| 590 |
+
"step": 97
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 1.46,
|
| 594 |
+
"learning_rate": 0.00021940197057092962,
|
| 595 |
+
"loss": 0.8197,
|
| 596 |
+
"step": 98
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 1.48,
|
| 600 |
+
"learning_rate": 0.00021617608807228086,
|
| 601 |
+
"loss": 0.8234,
|
| 602 |
+
"step": 99
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"epoch": 1.49,
|
| 606 |
+
"learning_rate": 0.00021294596367488718,
|
| 607 |
+
"loss": 0.8038,
|
| 608 |
+
"step": 100
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 1.51,
|
| 612 |
+
"learning_rate": 0.00020971244442288968,
|
| 613 |
+
"loss": 0.8075,
|
| 614 |
+
"step": 101
|
| 615 |
+
},
|
| 616 |
+
{
|
| 617 |
+
"epoch": 1.52,
|
| 618 |
+
"learning_rate": 0.00020647637825067122,
|
| 619 |
+
"loss": 0.7947,
|
| 620 |
+
"step": 102
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 1.54,
|
| 624 |
+
"learning_rate": 0.00020323861376050033,
|
| 625 |
+
"loss": 0.7924,
|
| 626 |
+
"step": 103
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 1.55,
|
| 630 |
+
"learning_rate": 0.0002,
|
| 631 |
+
"loss": 0.7874,
|
| 632 |
+
"step": 104
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 1.57,
|
| 636 |
+
"learning_rate": 0.00019676138623949972,
|
| 637 |
+
"loss": 0.8004,
|
| 638 |
+
"step": 105
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 1.58,
|
| 642 |
+
"learning_rate": 0.00019352362174932889,
|
| 643 |
+
"loss": 0.7932,
|
| 644 |
+
"step": 106
|
| 645 |
+
},
|
| 646 |
+
{
|
| 647 |
+
"epoch": 1.6,
|
| 648 |
+
"learning_rate": 0.00019028755557711042,
|
| 649 |
+
"loss": 0.7869,
|
| 650 |
+
"step": 107
|
| 651 |
+
},
|
| 652 |
+
{
|
| 653 |
+
"epoch": 1.61,
|
| 654 |
+
"learning_rate": 0.00018705403632511287,
|
| 655 |
+
"loss": 0.7622,
|
| 656 |
+
"step": 108
|
| 657 |
+
},
|
| 658 |
+
{
|
| 659 |
+
"epoch": 1.63,
|
| 660 |
+
"learning_rate": 0.0001838239119277192,
|
| 661 |
+
"loss": 0.7618,
|
| 662 |
+
"step": 109
|
| 663 |
+
},
|
| 664 |
+
{
|
| 665 |
+
"epoch": 1.64,
|
| 666 |
+
"learning_rate": 0.0001805980294290704,
|
| 667 |
+
"loss": 0.7826,
|
| 668 |
+
"step": 110
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 1.66,
|
| 672 |
+
"learning_rate": 0.00017737723476094315,
|
| 673 |
+
"loss": 0.7971,
|
| 674 |
+
"step": 111
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 1.67,
|
| 678 |
+
"learning_rate": 0.0001741623725209188,
|
| 679 |
+
"loss": 0.8068,
|
| 680 |
+
"step": 112
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 1.69,
|
| 684 |
+
"learning_rate": 0.000170954285750902,
|
| 685 |
+
"loss": 0.7831,
|
| 686 |
+
"step": 113
|
| 687 |
+
},
|
| 688 |
+
{
|
| 689 |
+
"epoch": 1.7,
|
| 690 |
+
"learning_rate": 0.00016775381571604804,
|
| 691 |
+
"loss": 0.7988,
|
| 692 |
+
"step": 114
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 1.72,
|
| 696 |
+
"learning_rate": 0.00016456180168415547,
|
| 697 |
+
"loss": 0.7912,
|
| 698 |
+
"step": 115
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 1.73,
|
| 702 |
+
"learning_rate": 0.00016137908070558242,
|
| 703 |
+
"loss": 0.7737,
|
| 704 |
+
"step": 116
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"epoch": 1.75,
|
| 708 |
+
"learning_rate": 0.000158206487393745,
|
| 709 |
+
"loss": 0.8103,
|
| 710 |
+
"step": 117
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 1.76,
|
| 714 |
+
"learning_rate": 0.0001550448537062542,
|
| 715 |
+
"loss": 0.7843,
|
| 716 |
+
"step": 118
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 1.78,
|
| 720 |
+
"learning_rate": 0.00015189500872674934,
|
| 721 |
+
"loss": 0.7957,
|
| 722 |
+
"step": 119
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 1.79,
|
| 726 |
+
"learning_rate": 0.00014875777844748552,
|
| 727 |
+
"loss": 0.7912,
|
| 728 |
+
"step": 120
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"epoch": 1.81,
|
| 732 |
+
"learning_rate": 0.00014563398555273143,
|
| 733 |
+
"loss": 0.79,
|
| 734 |
+
"step": 121
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 1.82,
|
| 738 |
+
"learning_rate": 0.00014252444920303438,
|
| 739 |
+
"loss": 0.8032,
|
| 740 |
+
"step": 122
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 1.84,
|
| 744 |
+
"learning_rate": 0.0001394299848204099,
|
| 745 |
+
"loss": 0.8148,
|
| 746 |
+
"step": 123
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 1.85,
|
| 750 |
+
"learning_rate": 0.00013635140387451128,
|
| 751 |
+
"loss": 0.7912,
|
| 752 |
+
"step": 124
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 1.87,
|
| 756 |
+
"learning_rate": 0.00013328951366983592,
|
| 757 |
+
"loss": 0.7999,
|
| 758 |
+
"step": 125
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 1.88,
|
| 762 |
+
"learning_rate": 0.00013024511713402353,
|
| 763 |
+
"loss": 0.7916,
|
| 764 |
+
"step": 126
|
| 765 |
+
},
|
| 766 |
+
{
|
| 767 |
+
"epoch": 1.9,
|
| 768 |
+
"learning_rate": 0.00012721901260730252,
|
| 769 |
+
"loss": 0.7848,
|
| 770 |
+
"step": 127
|
| 771 |
+
},
|
| 772 |
+
{
|
| 773 |
+
"epoch": 1.91,
|
| 774 |
+
"learning_rate": 0.00012421199363313865,
|
| 775 |
+
"loss": 0.7869,
|
| 776 |
+
"step": 128
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 1.93,
|
| 780 |
+
"learning_rate": 0.00012122484875014261,
|
| 781 |
+
"loss": 0.7495,
|
| 782 |
+
"step": 129
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 1.94,
|
| 786 |
+
"learning_rate": 0.00011825836128528883,
|
| 787 |
+
"loss": 0.761,
|
| 788 |
+
"step": 130
|
| 789 |
+
},
|
| 790 |
+
{
|
| 791 |
+
"epoch": 1.96,
|
| 792 |
+
"learning_rate": 0.00011531330914850204,
|
| 793 |
+
"loss": 0.7943,
|
| 794 |
+
"step": 131
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 1.97,
|
| 798 |
+
"learning_rate": 0.00011239046462866353,
|
| 799 |
+
"loss": 0.8015,
|
| 800 |
+
"step": 132
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 1.99,
|
| 804 |
+
"learning_rate": 0.00010949059419109224,
|
| 805 |
+
"loss": 0.7766,
|
| 806 |
+
"step": 133
|
| 807 |
+
},
|
| 808 |
+
{
|
| 809 |
+
"epoch": 2.0,
|
| 810 |
+
"learning_rate": 0.00010661445827655187,
|
| 811 |
+
"loss": 0.7663,
|
| 812 |
+
"step": 134
|
| 813 |
+
},
|
| 814 |
+
{
|
| 815 |
+
"epoch": 2.01,
|
| 816 |
+
"learning_rate": 0.00010376281110183899,
|
| 817 |
+
"loss": 0.5708,
|
| 818 |
+
"step": 135
|
| 819 |
+
},
|
| 820 |
+
{
|
| 821 |
+
"epoch": 2.03,
|
| 822 |
+
"learning_rate": 0.00010093640046200257,
|
| 823 |
+
"loss": 0.5661,
|
| 824 |
+
"step": 136
|
| 825 |
+
},
|
| 826 |
+
{
|
| 827 |
+
"epoch": 2.04,
|
| 828 |
+
"learning_rate": 9.813596753424747e-05,
|
| 829 |
+
"loss": 0.5428,
|
| 830 |
+
"step": 137
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 2.06,
|
| 834 |
+
"learning_rate": 9.536224668357398e-05,
|
| 835 |
+
"loss": 0.5465,
|
| 836 |
+
"step": 138
|
| 837 |
+
},
|
| 838 |
+
{
|
| 839 |
+
"epoch": 2.07,
|
| 840 |
+
"learning_rate": 9.261596527020324e-05,
|
| 841 |
+
"loss": 0.5272,
|
| 842 |
+
"step": 139
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 2.09,
|
| 846 |
+
"learning_rate": 8.98978434588393e-05,
|
| 847 |
+
"loss": 0.5375,
|
| 848 |
+
"step": 140
|
| 849 |
+
},
|
| 850 |
+
{
|
| 851 |
+
"epoch": 2.1,
|
| 852 |
+
"learning_rate": 8.720859402981868e-05,
|
| 853 |
+
"loss": 0.5253,
|
| 854 |
+
"step": 141
|
| 855 |
+
},
|
| 856 |
+
{
|
| 857 |
+
"epoch": 2.12,
|
| 858 |
+
"learning_rate": 8.454892219219617e-05,
|
| 859 |
+
"loss": 0.5265,
|
| 860 |
+
"step": 142
|
| 861 |
+
},
|
| 862 |
+
{
|
| 863 |
+
"epoch": 2.13,
|
| 864 |
+
"learning_rate": 8.191952539881553e-05,
|
| 865 |
+
"loss": 0.4915,
|
| 866 |
+
"step": 143
|
| 867 |
+
},
|
| 868 |
+
{
|
| 869 |
+
"epoch": 2.15,
|
| 870 |
+
"learning_rate": 7.932109316341507e-05,
|
| 871 |
+
"loss": 0.519,
|
| 872 |
+
"step": 144
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"epoch": 2.16,
|
| 876 |
+
"learning_rate": 7.675430687981453e-05,
|
| 877 |
+
"loss": 0.5061,
|
| 878 |
+
"step": 145
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 2.18,
|
| 882 |
+
"learning_rate": 7.42198396432311e-05,
|
| 883 |
+
"loss": 0.5109,
|
| 884 |
+
"step": 146
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 2.19,
|
| 888 |
+
"learning_rate": 7.171835607377206e-05,
|
| 889 |
+
"loss": 0.5141,
|
| 890 |
+
"step": 147
|
| 891 |
+
},
|
| 892 |
+
{
|
| 893 |
+
"epoch": 2.21,
|
| 894 |
+
"learning_rate": 6.925051214214955e-05,
|
| 895 |
+
"loss": 0.4993,
|
| 896 |
+
"step": 148
|
| 897 |
+
},
|
| 898 |
+
{
|
| 899 |
+
"epoch": 2.22,
|
| 900 |
+
"learning_rate": 6.681695499766383e-05,
|
| 901 |
+
"loss": 0.4964,
|
| 902 |
+
"step": 149
|
| 903 |
+
},
|
| 904 |
+
{
|
| 905 |
+
"epoch": 2.24,
|
| 906 |
+
"learning_rate": 6.441832279849935e-05,
|
| 907 |
+
"loss": 0.4871,
|
| 908 |
+
"step": 150
|
| 909 |
+
},
|
| 910 |
+
{
|
| 911 |
+
"epoch": 2.25,
|
| 912 |
+
"learning_rate": 6.205524454437915e-05,
|
| 913 |
+
"loss": 0.4901,
|
| 914 |
+
"step": 151
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 2.27,
|
| 918 |
+
"learning_rate": 5.972833991162017e-05,
|
| 919 |
+
"loss": 0.4883,
|
| 920 |
+
"step": 152
|
| 921 |
+
},
|
| 922 |
+
{
|
| 923 |
+
"epoch": 2.28,
|
| 924 |
+
"learning_rate": 5.7438219090634206e-05,
|
| 925 |
+
"loss": 0.4976,
|
| 926 |
+
"step": 153
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 2.3,
|
| 930 |
+
"learning_rate": 5.518548262591574e-05,
|
| 931 |
+
"loss": 0.5058,
|
| 932 |
+
"step": 154
|
| 933 |
+
},
|
| 934 |
+
{
|
| 935 |
+
"epoch": 2.31,
|
| 936 |
+
"learning_rate": 5.297072125855997e-05,
|
| 937 |
+
"loss": 0.4878,
|
| 938 |
+
"step": 155
|
| 939 |
+
},
|
| 940 |
+
{
|
| 941 |
+
"epoch": 2.33,
|
| 942 |
+
"learning_rate": 5.0794515771350795e-05,
|
| 943 |
+
"loss": 0.4967,
|
| 944 |
+
"step": 156
|
| 945 |
+
},
|
| 946 |
+
{
|
| 947 |
+
"epoch": 2.34,
|
| 948 |
+
"learning_rate": 4.865743683646094e-05,
|
| 949 |
+
"loss": 0.478,
|
| 950 |
+
"step": 157
|
| 951 |
+
},
|
| 952 |
+
{
|
| 953 |
+
"epoch": 2.36,
|
| 954 |
+
"learning_rate": 4.6560044865802766e-05,
|
| 955 |
+
"loss": 0.5068,
|
| 956 |
+
"step": 158
|
| 957 |
+
},
|
| 958 |
+
{
|
| 959 |
+
"epoch": 2.37,
|
| 960 |
+
"learning_rate": 4.450288986407019e-05,
|
| 961 |
+
"loss": 0.4856,
|
| 962 |
+
"step": 159
|
| 963 |
+
},
|
| 964 |
+
{
|
| 965 |
+
"epoch": 2.39,
|
| 966 |
+
"learning_rate": 4.2486511284509155e-05,
|
| 967 |
+
"loss": 0.4929,
|
| 968 |
+
"step": 160
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 2.4,
|
| 972 |
+
"learning_rate": 4.051143788745588e-05,
|
| 973 |
+
"loss": 0.4845,
|
| 974 |
+
"step": 161
|
| 975 |
+
},
|
| 976 |
+
{
|
| 977 |
+
"epoch": 2.42,
|
| 978 |
+
"learning_rate": 3.857818760167813e-05,
|
| 979 |
+
"loss": 0.5001,
|
| 980 |
+
"step": 162
|
| 981 |
+
},
|
| 982 |
+
{
|
| 983 |
+
"epoch": 2.43,
|
| 984 |
+
"learning_rate": 3.668726738855779e-05,
|
| 985 |
+
"loss": 0.4841,
|
| 986 |
+
"step": 163
|
| 987 |
+
},
|
| 988 |
+
{
|
| 989 |
+
"epoch": 2.45,
|
| 990 |
+
"learning_rate": 3.483917310914872e-05,
|
| 991 |
+
"loss": 0.4843,
|
| 992 |
+
"step": 164
|
| 993 |
+
},
|
| 994 |
+
{
|
| 995 |
+
"epoch": 2.46,
|
| 996 |
+
"learning_rate": 3.30343893941458e-05,
|
| 997 |
+
"loss": 0.4976,
|
| 998 |
+
"step": 165
|
| 999 |
+
},
|
| 1000 |
+
{
|
| 1001 |
+
"epoch": 2.48,
|
| 1002 |
+
"learning_rate": 3.127338951679912e-05,
|
| 1003 |
+
"loss": 0.4971,
|
| 1004 |
+
"step": 166
|
| 1005 |
+
},
|
| 1006 |
+
{
|
| 1007 |
+
"epoch": 2.49,
|
| 1008 |
+
"learning_rate": 2.9556635268806167e-05,
|
| 1009 |
+
"loss": 0.5043,
|
| 1010 |
+
"step": 167
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 2.51,
|
| 1014 |
+
"learning_rate": 2.788457683921495e-05,
|
| 1015 |
+
"loss": 0.4812,
|
| 1016 |
+
"step": 168
|
| 1017 |
+
},
|
| 1018 |
+
{
|
| 1019 |
+
"epoch": 2.52,
|
| 1020 |
+
"learning_rate": 2.6257652696369774e-05,
|
| 1021 |
+
"loss": 0.4702,
|
| 1022 |
+
"step": 169
|
| 1023 |
+
},
|
| 1024 |
+
{
|
| 1025 |
+
"epoch": 2.54,
|
| 1026 |
+
"learning_rate": 2.467628947293048e-05,
|
| 1027 |
+
"loss": 0.5025,
|
| 1028 |
+
"step": 170
|
| 1029 |
+
},
|
| 1030 |
+
{
|
| 1031 |
+
"epoch": 2.55,
|
| 1032 |
+
"learning_rate": 2.314090185399531e-05,
|
| 1033 |
+
"loss": 0.4985,
|
| 1034 |
+
"step": 171
|
| 1035 |
+
},
|
| 1036 |
+
{
|
| 1037 |
+
"epoch": 2.57,
|
| 1038 |
+
"learning_rate": 2.1651892468357148e-05,
|
| 1039 |
+
"loss": 0.5118,
|
| 1040 |
+
"step": 172
|
| 1041 |
+
},
|
| 1042 |
+
{
|
| 1043 |
+
"epoch": 2.58,
|
| 1044 |
+
"learning_rate": 2.020965178292096e-05,
|
| 1045 |
+
"loss": 0.489,
|
| 1046 |
+
"step": 173
|
| 1047 |
+
},
|
| 1048 |
+
{
|
| 1049 |
+
"epoch": 2.6,
|
| 1050 |
+
"learning_rate": 1.8814558000310623e-05,
|
| 1051 |
+
"loss": 0.4698,
|
| 1052 |
+
"step": 174
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 2.61,
|
| 1056 |
+
"learning_rate": 1.7466976959691993e-05,
|
| 1057 |
+
"loss": 0.4908,
|
| 1058 |
+
"step": 175
|
| 1059 |
+
},
|
| 1060 |
+
{
|
| 1061 |
+
"epoch": 2.63,
|
| 1062 |
+
"learning_rate": 1.6167262040837582e-05,
|
| 1063 |
+
"loss": 0.4752,
|
| 1064 |
+
"step": 176
|
| 1065 |
+
},
|
| 1066 |
+
{
|
| 1067 |
+
"epoch": 2.64,
|
| 1068 |
+
"learning_rate": 1.4915754071459175e-05,
|
| 1069 |
+
"loss": 0.4993,
|
| 1070 |
+
"step": 177
|
| 1071 |
+
},
|
| 1072 |
+
{
|
| 1073 |
+
"epoch": 2.66,
|
| 1074 |
+
"learning_rate": 1.371278123783155e-05,
|
| 1075 |
+
"loss": 0.4926,
|
| 1076 |
+
"step": 178
|
| 1077 |
+
},
|
| 1078 |
+
{
|
| 1079 |
+
"epoch": 2.67,
|
| 1080 |
+
"learning_rate": 1.2558658998731299e-05,
|
| 1081 |
+
"loss": 0.4845,
|
| 1082 |
+
"step": 179
|
| 1083 |
+
},
|
| 1084 |
+
{
|
| 1085 |
+
"epoch": 2.69,
|
| 1086 |
+
"learning_rate": 1.1453690002713146e-05,
|
| 1087 |
+
"loss": 0.4724,
|
| 1088 |
+
"step": 180
|
| 1089 |
+
},
|
| 1090 |
+
{
|
| 1091 |
+
"epoch": 2.7,
|
| 1092 |
+
"learning_rate": 1.0398164008745915e-05,
|
| 1093 |
+
"loss": 0.4993,
|
| 1094 |
+
"step": 181
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 2.72,
|
| 1098 |
+
"learning_rate": 9.392357810227825e-06,
|
| 1099 |
+
"loss": 0.5016,
|
| 1100 |
+
"step": 182
|
| 1101 |
+
},
|
| 1102 |
+
{
|
| 1103 |
+
"epoch": 2.73,
|
| 1104 |
+
"learning_rate": 8.436535162402593e-06,
|
| 1105 |
+
"loss": 0.4747,
|
| 1106 |
+
"step": 183
|
| 1107 |
+
},
|
| 1108 |
+
{
|
| 1109 |
+
"epoch": 2.75,
|
| 1110 |
+
"learning_rate": 7.530946713193965e-06,
|
| 1111 |
+
"loss": 0.4979,
|
| 1112 |
+
"step": 184
|
| 1113 |
+
},
|
| 1114 |
+
{
|
| 1115 |
+
"epoch": 2.76,
|
| 1116 |
+
"learning_rate": 6.675829937477773e-06,
|
| 1117 |
+
"loss": 0.4925,
|
| 1118 |
+
"step": 185
|
| 1119 |
+
},
|
| 1120 |
+
{
|
| 1121 |
+
"epoch": 2.78,
|
| 1122 |
+
"learning_rate": 5.871409074808165e-06,
|
| 1123 |
+
"loss": 0.4798,
|
| 1124 |
+
"step": 186
|
| 1125 |
+
},
|
| 1126 |
+
{
|
| 1127 |
+
"epoch": 2.79,
|
| 1128 |
+
"learning_rate": 5.1178950706147975e-06,
|
| 1129 |
+
"loss": 0.4896,
|
| 1130 |
+
"step": 187
|
| 1131 |
+
},
|
| 1132 |
+
{
|
| 1133 |
+
"epoch": 2.81,
|
| 1134 |
+
"learning_rate": 4.415485520885887e-06,
|
| 1135 |
+
"loss": 0.4885,
|
| 1136 |
+
"step": 188
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 2.82,
|
| 1140 |
+
"learning_rate": 3.76436462035219e-06,
|
| 1141 |
+
"loss": 0.4882,
|
| 1142 |
+
"step": 189
|
| 1143 |
+
},
|
| 1144 |
+
{
|
| 1145 |
+
"epoch": 2.84,
|
| 1146 |
+
"learning_rate": 3.1647031141851525e-06,
|
| 1147 |
+
"loss": 0.4781,
|
| 1148 |
+
"step": 190
|
| 1149 |
+
},
|
| 1150 |
+
{
|
| 1151 |
+
"epoch": 2.85,
|
| 1152 |
+
"learning_rate": 2.61665825322186e-06,
|
| 1153 |
+
"loss": 0.4995,
|
| 1154 |
+
"step": 191
|
| 1155 |
+
},
|
| 1156 |
+
{
|
| 1157 |
+
"epoch": 2.87,
|
| 1158 |
+
"learning_rate": 2.120373752728799e-06,
|
| 1159 |
+
"loss": 0.4884,
|
| 1160 |
+
"step": 192
|
| 1161 |
+
},
|
| 1162 |
+
{
|
| 1163 |
+
"epoch": 2.88,
|
| 1164 |
+
"learning_rate": 1.6759797547149847e-06,
|
| 1165 |
+
"loss": 0.493,
|
| 1166 |
+
"step": 193
|
| 1167 |
+
},
|
| 1168 |
+
{
|
| 1169 |
+
"epoch": 2.9,
|
| 1170 |
+
"learning_rate": 1.283592793804478e-06,
|
| 1171 |
+
"loss": 0.4934,
|
| 1172 |
+
"step": 194
|
| 1173 |
+
},
|
| 1174 |
+
{
|
| 1175 |
+
"epoch": 2.91,
|
| 1176 |
+
"learning_rate": 9.433157666772107e-07,
|
| 1177 |
+
"loss": 0.4861,
|
| 1178 |
+
"step": 195
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 2.93,
|
| 1182 |
+
"learning_rate": 6.552379050861257e-07,
|
| 1183 |
+
"loss": 0.4713,
|
| 1184 |
+
"step": 196
|
| 1185 |
+
},
|
| 1186 |
+
{
|
| 1187 |
+
"epoch": 2.94,
|
| 1188 |
+
"learning_rate": 4.194347524576703e-07,
|
| 1189 |
+
"loss": 0.4767,
|
| 1190 |
+
"step": 197
|
| 1191 |
+
},
|
| 1192 |
+
{
|
| 1193 |
+
"epoch": 2.96,
|
| 1194 |
+
"learning_rate": 2.359681440818662e-07,
|
| 1195 |
+
"loss": 0.4881,
|
| 1196 |
+
"step": 198
|
| 1197 |
+
},
|
| 1198 |
+
{
|
| 1199 |
+
"epoch": 2.97,
|
| 1200 |
+
"learning_rate": 1.0488619089701335e-07,
|
| 1201 |
+
"loss": 0.4859,
|
| 1202 |
+
"step": 199
|
| 1203 |
+
},
|
| 1204 |
+
{
|
| 1205 |
+
"epoch": 2.99,
|
| 1206 |
+
"learning_rate": 2.6223266873559582e-08,
|
| 1207 |
+
"loss": 0.4896,
|
| 1208 |
+
"step": 200
|
| 1209 |
+
},
|
| 1210 |
+
{
|
| 1211 |
+
"epoch": 3.0,
|
| 1212 |
+
"learning_rate": 0.0,
|
| 1213 |
+
"loss": 0.4539,
|
| 1214 |
+
"step": 201
|
| 1215 |
+
},
|
| 1216 |
+
{
|
| 1217 |
+
"epoch": 3.0,
|
| 1218 |
+
"step": 201,
|
| 1219 |
+
"total_flos": 483875872178176.0,
|
| 1220 |
+
"train_loss": 0.7875154921367987,
|
| 1221 |
+
"train_runtime": 6065.9891,
|
| 1222 |
+
"train_samples_per_second": 4.237,
|
| 1223 |
+
"train_steps_per_second": 0.033
|
| 1224 |
+
}
|
| 1225 |
+
],
|
| 1226 |
+
"max_steps": 201,
|
| 1227 |
+
"num_train_epochs": 3,
|
| 1228 |
+
"total_flos": 483875872178176.0,
|
| 1229 |
+
"trial_name": null,
|
| 1230 |
+
"trial_params": null
|
| 1231 |
+
}
|