End of training
Browse files- README.md +60 -0
- adapter_config.json +26 -0
- adapter_model.safetensors +3 -0
- config.json +49 -0
- model_args.json +14 -0
- non_lora_trainables.bin +3 -0
- special_tokens_map.json +24 -0
- tokenizer.model +3 -0
- tokenizer_config.json +44 -0
- trainer_state.json +2634 -0
- training_args.bin +3 -0
- training_args.json +141 -0
README.md
ADDED
|
@@ -0,0 +1,60 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
base_model: microsoft/llava-med-v1.5-mistral-7b
|
| 4 |
+
tags:
|
| 5 |
+
- generated_from_trainer
|
| 6 |
+
model-index:
|
| 7 |
+
- name: llava-med-spect-persona-ffr-20250409
|
| 8 |
+
results: []
|
| 9 |
+
library_name: peft
|
| 10 |
+
---
|
| 11 |
+
|
| 12 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 13 |
+
should probably proofread and complete it, then remove this comment. -->
|
| 14 |
+
|
| 15 |
+
# llava-med-spect-persona-ffr-20250409
|
| 16 |
+
|
| 17 |
+
This model is a fine-tuned version of [microsoft/llava-med-v1.5-mistral-7b](https://huggingface.co/microsoft/llava-med-v1.5-mistral-7b) on an unknown dataset.
|
| 18 |
+
|
| 19 |
+
## Model description
|
| 20 |
+
|
| 21 |
+
More information needed
|
| 22 |
+
|
| 23 |
+
## Intended uses & limitations
|
| 24 |
+
|
| 25 |
+
More information needed
|
| 26 |
+
|
| 27 |
+
## Training and evaluation data
|
| 28 |
+
|
| 29 |
+
More information needed
|
| 30 |
+
|
| 31 |
+
## Training procedure
|
| 32 |
+
|
| 33 |
+
### Training hyperparameters
|
| 34 |
+
|
| 35 |
+
The following hyperparameters were used during training:
|
| 36 |
+
- learning_rate: 0.0001
|
| 37 |
+
- train_batch_size: 4
|
| 38 |
+
- eval_batch_size: 2
|
| 39 |
+
- seed: 42
|
| 40 |
+
- distributed_type: multi-GPU
|
| 41 |
+
- num_devices: 4
|
| 42 |
+
- gradient_accumulation_steps: 4
|
| 43 |
+
- total_train_batch_size: 64
|
| 44 |
+
- total_eval_batch_size: 8
|
| 45 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 46 |
+
- lr_scheduler_type: cosine
|
| 47 |
+
- lr_scheduler_warmup_ratio: 0.02
|
| 48 |
+
- num_epochs: 1.0
|
| 49 |
+
- mixed_precision_training: Native AMP
|
| 50 |
+
|
| 51 |
+
### Training results
|
| 52 |
+
|
| 53 |
+
|
| 54 |
+
|
| 55 |
+
### Framework versions
|
| 56 |
+
|
| 57 |
+
- PEFT 0.4.0
|
| 58 |
+
- Transformers 4.37.2
|
| 59 |
+
- Pytorch 2.1.2+cu121
|
| 60 |
+
- Tokenizers 0.15.1
|
adapter_config.json
ADDED
|
@@ -0,0 +1,26 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"auto_mapping": null,
|
| 3 |
+
"base_model_name_or_path": "microsoft/llava-med-v1.5-mistral-7b",
|
| 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": 32,
|
| 11 |
+
"lora_dropout": 0.05,
|
| 12 |
+
"modules_to_save": null,
|
| 13 |
+
"peft_type": "LORA",
|
| 14 |
+
"r": 16,
|
| 15 |
+
"revision": null,
|
| 16 |
+
"target_modules": [
|
| 17 |
+
"v_proj",
|
| 18 |
+
"k_proj",
|
| 19 |
+
"gate_proj",
|
| 20 |
+
"o_proj",
|
| 21 |
+
"down_proj",
|
| 22 |
+
"up_proj",
|
| 23 |
+
"q_proj"
|
| 24 |
+
],
|
| 25 |
+
"task_type": "CAUSAL_LM"
|
| 26 |
+
}
|
adapter_model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1c72b657ca72ada314cf35b395edbf0d889717b22afc6d177d4dcf22d80659c3
|
| 3 |
+
size 167832240
|
config.json
ADDED
|
@@ -0,0 +1,49 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "microsoft/llava-med-v1.5-mistral-7b",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"LlavaMistralForCausalLM"
|
| 5 |
+
],
|
| 6 |
+
"attention_dropout": 0.0,
|
| 7 |
+
"bos_token_id": 1,
|
| 8 |
+
"eos_token_id": 2,
|
| 9 |
+
"feature_outs": "encoder+decoder",
|
| 10 |
+
"freeze_mm_mlp_adapter": false,
|
| 11 |
+
"hidden_act": "silu",
|
| 12 |
+
"hidden_size": 4096,
|
| 13 |
+
"image_aspect_ratio": "pad",
|
| 14 |
+
"img_size": 640,
|
| 15 |
+
"initializer_range": 0.02,
|
| 16 |
+
"intermediate_size": 14336,
|
| 17 |
+
"max_position_embeddings": 32768,
|
| 18 |
+
"mm_hidden_size": 1024,
|
| 19 |
+
"mm_patch_merge_type": "flat",
|
| 20 |
+
"mm_projector_lr": 1e-05,
|
| 21 |
+
"mm_projector_type": "mlp2x_gelu",
|
| 22 |
+
"mm_use_im_patch_token": false,
|
| 23 |
+
"mm_use_im_start_end": false,
|
| 24 |
+
"mm_vision_select_feature": "patch",
|
| 25 |
+
"mm_vision_select_layer": -2,
|
| 26 |
+
"mm_vision_tower": "microsoft/llava-med-v1.5-mistral-7b",
|
| 27 |
+
"model_type": "llava_mistral",
|
| 28 |
+
"num_attention_heads": 32,
|
| 29 |
+
"num_hidden_layers": 32,
|
| 30 |
+
"num_key_value_heads": 8,
|
| 31 |
+
"proj_vis_to_txt_tokens": false,
|
| 32 |
+
"prompt_segtok_w_instruct": false,
|
| 33 |
+
"rms_norm_eps": 1e-05,
|
| 34 |
+
"rope_theta": 1000000.0,
|
| 35 |
+
"segtok_posembed": "sincos",
|
| 36 |
+
"sliding_window": null,
|
| 37 |
+
"tie_word_embeddings": false,
|
| 38 |
+
"tokenizer_model_max_length": 2048,
|
| 39 |
+
"tokenizer_padding_side": "right",
|
| 40 |
+
"torch_dtype": "bfloat16",
|
| 41 |
+
"transformers_version": "4.37.2",
|
| 42 |
+
"tune_mm_mlp_adapter": false,
|
| 43 |
+
"tune_vision_tokenizer": "none",
|
| 44 |
+
"use_cache": true,
|
| 45 |
+
"use_mm_proj": true,
|
| 46 |
+
"vision_backbone": "convnextlarge",
|
| 47 |
+
"vision_tokenizer_lr": null,
|
| 48 |
+
"vocab_size": 32000
|
| 49 |
+
}
|
model_args.json
ADDED
|
@@ -0,0 +1,14 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"model_name_or_path": "microsoft/llava-med-v1.5-mistral-7b",
|
| 3 |
+
"version": "v1",
|
| 4 |
+
"freeze_backbone": false,
|
| 5 |
+
"tune_mm_mlp_adapter": false,
|
| 6 |
+
"vision_tower": "microsoft/llava-med-v1.5-mistral-7b",
|
| 7 |
+
"mm_vision_select_layer": -2,
|
| 8 |
+
"pretrain_mm_mlp_adapter": null,
|
| 9 |
+
"mm_projector_type": "mlp2x_gelu",
|
| 10 |
+
"mm_use_im_start_end": false,
|
| 11 |
+
"mm_use_im_patch_token": false,
|
| 12 |
+
"mm_patch_merge_type": "flat",
|
| 13 |
+
"mm_vision_select_feature": "patch"
|
| 14 |
+
}
|
non_lora_trainables.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c5f1709084e4975ae4f8b3b52b73b33a2d21259361bfd780844a4a5b7e864cc5
|
| 3 |
+
size 41961648
|
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": "<unk>",
|
| 17 |
+
"unk_token": {
|
| 18 |
+
"content": "<unk>",
|
| 19 |
+
"lstrip": false,
|
| 20 |
+
"normalized": false,
|
| 21 |
+
"rstrip": false,
|
| 22 |
+
"single_word": false
|
| 23 |
+
}
|
| 24 |
+
}
|
tokenizer.model
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:dadfd56d766715c61d2ef780a525ab43b8e6da4de6865bda3d95fdef5e134055
|
| 3 |
+
size 493443
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,44 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_bos_token": true,
|
| 3 |
+
"add_eos_token": false,
|
| 4 |
+
"added_tokens_decoder": {
|
| 5 |
+
"0": {
|
| 6 |
+
"content": "<unk>",
|
| 7 |
+
"lstrip": false,
|
| 8 |
+
"normalized": false,
|
| 9 |
+
"rstrip": false,
|
| 10 |
+
"single_word": false,
|
| 11 |
+
"special": true
|
| 12 |
+
},
|
| 13 |
+
"1": {
|
| 14 |
+
"content": "<s>",
|
| 15 |
+
"lstrip": false,
|
| 16 |
+
"normalized": false,
|
| 17 |
+
"rstrip": false,
|
| 18 |
+
"single_word": false,
|
| 19 |
+
"special": true
|
| 20 |
+
},
|
| 21 |
+
"2": {
|
| 22 |
+
"content": "</s>",
|
| 23 |
+
"lstrip": false,
|
| 24 |
+
"normalized": false,
|
| 25 |
+
"rstrip": false,
|
| 26 |
+
"single_word": false,
|
| 27 |
+
"special": true
|
| 28 |
+
}
|
| 29 |
+
},
|
| 30 |
+
"additional_special_tokens": [],
|
| 31 |
+
"bos_token": "<s>",
|
| 32 |
+
"chat_template": "{{ bos_token }}{% for message in messages %}{% if (message['role'] == 'user') != (loop.index0 % 2 == 0) %}{{ raise_exception('Conversation roles must alternate user/assistant/user/assistant/...') }}{% endif %}{% if message['role'] == 'user' %}{{ '[INST] ' + message['content'] + ' [/INST]' }}{% elif message['role'] == 'assistant' %}{{ message['content'] + eos_token}}{% else %}{{ raise_exception('Only user and assistant roles are supported!') }}{% endif %}{% endfor %}",
|
| 33 |
+
"clean_up_tokenization_spaces": false,
|
| 34 |
+
"eos_token": "</s>",
|
| 35 |
+
"legacy": true,
|
| 36 |
+
"model_max_length": 2048,
|
| 37 |
+
"pad_token": "<unk>",
|
| 38 |
+
"padding_side": "right",
|
| 39 |
+
"sp_model_kwargs": {},
|
| 40 |
+
"spaces_between_special_tokens": false,
|
| 41 |
+
"tokenizer_class": "LlamaTokenizer",
|
| 42 |
+
"unk_token": "<unk>",
|
| 43 |
+
"use_default_system_prompt": false
|
| 44 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,2634 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 1.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 434,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.0,
|
| 13 |
+
"learning_rate": 1.1111111111111112e-05,
|
| 14 |
+
"loss": 2.3254,
|
| 15 |
+
"step": 1
|
| 16 |
+
},
|
| 17 |
+
{
|
| 18 |
+
"epoch": 0.0,
|
| 19 |
+
"learning_rate": 2.2222222222222223e-05,
|
| 20 |
+
"loss": 2.4112,
|
| 21 |
+
"step": 2
|
| 22 |
+
},
|
| 23 |
+
{
|
| 24 |
+
"epoch": 0.01,
|
| 25 |
+
"learning_rate": 3.3333333333333335e-05,
|
| 26 |
+
"loss": 2.3104,
|
| 27 |
+
"step": 3
|
| 28 |
+
},
|
| 29 |
+
{
|
| 30 |
+
"epoch": 0.01,
|
| 31 |
+
"learning_rate": 4.4444444444444447e-05,
|
| 32 |
+
"loss": 2.1862,
|
| 33 |
+
"step": 4
|
| 34 |
+
},
|
| 35 |
+
{
|
| 36 |
+
"epoch": 0.01,
|
| 37 |
+
"learning_rate": 5.555555555555556e-05,
|
| 38 |
+
"loss": 1.8671,
|
| 39 |
+
"step": 5
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.01,
|
| 43 |
+
"learning_rate": 6.666666666666667e-05,
|
| 44 |
+
"loss": 1.632,
|
| 45 |
+
"step": 6
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.02,
|
| 49 |
+
"learning_rate": 7.777777777777778e-05,
|
| 50 |
+
"loss": 1.2926,
|
| 51 |
+
"step": 7
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.02,
|
| 55 |
+
"learning_rate": 8.888888888888889e-05,
|
| 56 |
+
"loss": 1.1784,
|
| 57 |
+
"step": 8
|
| 58 |
+
},
|
| 59 |
+
{
|
| 60 |
+
"epoch": 0.02,
|
| 61 |
+
"learning_rate": 0.0001,
|
| 62 |
+
"loss": 0.9747,
|
| 63 |
+
"step": 9
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 0.02,
|
| 67 |
+
"learning_rate": 9.999863397100895e-05,
|
| 68 |
+
"loss": 0.6938,
|
| 69 |
+
"step": 10
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 0.03,
|
| 73 |
+
"learning_rate": 9.999453595867717e-05,
|
| 74 |
+
"loss": 0.4035,
|
| 75 |
+
"step": 11
|
| 76 |
+
},
|
| 77 |
+
{
|
| 78 |
+
"epoch": 0.03,
|
| 79 |
+
"learning_rate": 9.998770618692483e-05,
|
| 80 |
+
"loss": 0.228,
|
| 81 |
+
"step": 12
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"epoch": 0.03,
|
| 85 |
+
"learning_rate": 9.997814502893856e-05,
|
| 86 |
+
"loss": 0.202,
|
| 87 |
+
"step": 13
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"epoch": 0.03,
|
| 91 |
+
"learning_rate": 9.996585300715116e-05,
|
| 92 |
+
"loss": 0.1499,
|
| 93 |
+
"step": 14
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.03,
|
| 97 |
+
"learning_rate": 9.99508307932129e-05,
|
| 98 |
+
"loss": 0.0606,
|
| 99 |
+
"step": 15
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"epoch": 0.04,
|
| 103 |
+
"learning_rate": 9.993307920795501e-05,
|
| 104 |
+
"loss": 0.0492,
|
| 105 |
+
"step": 16
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 0.04,
|
| 109 |
+
"learning_rate": 9.991259922134466e-05,
|
| 110 |
+
"loss": 0.0409,
|
| 111 |
+
"step": 17
|
| 112 |
+
},
|
| 113 |
+
{
|
| 114 |
+
"epoch": 0.04,
|
| 115 |
+
"learning_rate": 9.98893919524321e-05,
|
| 116 |
+
"loss": 0.0374,
|
| 117 |
+
"step": 18
|
| 118 |
+
},
|
| 119 |
+
{
|
| 120 |
+
"epoch": 0.04,
|
| 121 |
+
"learning_rate": 9.986345866928941e-05,
|
| 122 |
+
"loss": 0.0137,
|
| 123 |
+
"step": 19
|
| 124 |
+
},
|
| 125 |
+
{
|
| 126 |
+
"epoch": 0.05,
|
| 127 |
+
"learning_rate": 9.983480078894124e-05,
|
| 128 |
+
"loss": 0.0101,
|
| 129 |
+
"step": 20
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.05,
|
| 133 |
+
"learning_rate": 9.98034198772874e-05,
|
| 134 |
+
"loss": 0.0066,
|
| 135 |
+
"step": 21
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.05,
|
| 139 |
+
"learning_rate": 9.976931764901733e-05,
|
| 140 |
+
"loss": 0.0057,
|
| 141 |
+
"step": 22
|
| 142 |
+
},
|
| 143 |
+
{
|
| 144 |
+
"epoch": 0.05,
|
| 145 |
+
"learning_rate": 9.973249596751631e-05,
|
| 146 |
+
"loss": 0.0359,
|
| 147 |
+
"step": 23
|
| 148 |
+
},
|
| 149 |
+
{
|
| 150 |
+
"epoch": 0.06,
|
| 151 |
+
"learning_rate": 9.96929568447637e-05,
|
| 152 |
+
"loss": 0.0038,
|
| 153 |
+
"step": 24
|
| 154 |
+
},
|
| 155 |
+
{
|
| 156 |
+
"epoch": 0.06,
|
| 157 |
+
"learning_rate": 9.965070244122301e-05,
|
| 158 |
+
"loss": 0.0014,
|
| 159 |
+
"step": 25
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 0.06,
|
| 163 |
+
"learning_rate": 9.96057350657239e-05,
|
| 164 |
+
"loss": 0.0004,
|
| 165 |
+
"step": 26
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 0.06,
|
| 169 |
+
"learning_rate": 9.955805717533586e-05,
|
| 170 |
+
"loss": 0.0017,
|
| 171 |
+
"step": 27
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 0.06,
|
| 175 |
+
"learning_rate": 9.950767137523416e-05,
|
| 176 |
+
"loss": 0.0003,
|
| 177 |
+
"step": 28
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.07,
|
| 181 |
+
"learning_rate": 9.94545804185573e-05,
|
| 182 |
+
"loss": 0.0002,
|
| 183 |
+
"step": 29
|
| 184 |
+
},
|
| 185 |
+
{
|
| 186 |
+
"epoch": 0.07,
|
| 187 |
+
"learning_rate": 9.939878720625675e-05,
|
| 188 |
+
"loss": 0.0002,
|
| 189 |
+
"step": 30
|
| 190 |
+
},
|
| 191 |
+
{
|
| 192 |
+
"epoch": 0.07,
|
| 193 |
+
"learning_rate": 9.934029478693832e-05,
|
| 194 |
+
"loss": 0.0002,
|
| 195 |
+
"step": 31
|
| 196 |
+
},
|
| 197 |
+
{
|
| 198 |
+
"epoch": 0.07,
|
| 199 |
+
"learning_rate": 9.927910635669562e-05,
|
| 200 |
+
"loss": 0.0001,
|
| 201 |
+
"step": 32
|
| 202 |
+
},
|
| 203 |
+
{
|
| 204 |
+
"epoch": 0.08,
|
| 205 |
+
"learning_rate": 9.921522525893548e-05,
|
| 206 |
+
"loss": 0.0337,
|
| 207 |
+
"step": 33
|
| 208 |
+
},
|
| 209 |
+
{
|
| 210 |
+
"epoch": 0.08,
|
| 211 |
+
"learning_rate": 9.91486549841951e-05,
|
| 212 |
+
"loss": 0.0001,
|
| 213 |
+
"step": 34
|
| 214 |
+
},
|
| 215 |
+
{
|
| 216 |
+
"epoch": 0.08,
|
| 217 |
+
"learning_rate": 9.907939916995152e-05,
|
| 218 |
+
"loss": 0.0032,
|
| 219 |
+
"step": 35
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.08,
|
| 223 |
+
"learning_rate": 9.900746160042273e-05,
|
| 224 |
+
"loss": 0.0001,
|
| 225 |
+
"step": 36
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 0.09,
|
| 229 |
+
"learning_rate": 9.893284620636099e-05,
|
| 230 |
+
"loss": 0.0002,
|
| 231 |
+
"step": 37
|
| 232 |
+
},
|
| 233 |
+
{
|
| 234 |
+
"epoch": 0.09,
|
| 235 |
+
"learning_rate": 9.885555706483792e-05,
|
| 236 |
+
"loss": 0.0005,
|
| 237 |
+
"step": 38
|
| 238 |
+
},
|
| 239 |
+
{
|
| 240 |
+
"epoch": 0.09,
|
| 241 |
+
"learning_rate": 9.877559839902184e-05,
|
| 242 |
+
"loss": 0.0006,
|
| 243 |
+
"step": 39
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"epoch": 0.09,
|
| 247 |
+
"learning_rate": 9.869297457794699e-05,
|
| 248 |
+
"loss": 0.0001,
|
| 249 |
+
"step": 40
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 0.09,
|
| 253 |
+
"learning_rate": 9.860769011627475e-05,
|
| 254 |
+
"loss": 0.0002,
|
| 255 |
+
"step": 41
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 0.1,
|
| 259 |
+
"learning_rate": 9.851974967404704e-05,
|
| 260 |
+
"loss": 0.0001,
|
| 261 |
+
"step": 42
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.1,
|
| 265 |
+
"learning_rate": 9.842915805643155e-05,
|
| 266 |
+
"loss": 0.0001,
|
| 267 |
+
"step": 43
|
| 268 |
+
},
|
| 269 |
+
{
|
| 270 |
+
"epoch": 0.1,
|
| 271 |
+
"learning_rate": 9.833592021345937e-05,
|
| 272 |
+
"loss": 0.0001,
|
| 273 |
+
"step": 44
|
| 274 |
+
},
|
| 275 |
+
{
|
| 276 |
+
"epoch": 0.1,
|
| 277 |
+
"learning_rate": 9.824004123975435e-05,
|
| 278 |
+
"loss": 0.0001,
|
| 279 |
+
"step": 45
|
| 280 |
+
},
|
| 281 |
+
{
|
| 282 |
+
"epoch": 0.11,
|
| 283 |
+
"learning_rate": 9.814152637425478e-05,
|
| 284 |
+
"loss": 0.0001,
|
| 285 |
+
"step": 46
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 0.11,
|
| 289 |
+
"learning_rate": 9.804038099992717e-05,
|
| 290 |
+
"loss": 0.0001,
|
| 291 |
+
"step": 47
|
| 292 |
+
},
|
| 293 |
+
{
|
| 294 |
+
"epoch": 0.11,
|
| 295 |
+
"learning_rate": 9.793661064347205e-05,
|
| 296 |
+
"loss": 0.0001,
|
| 297 |
+
"step": 48
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 0.11,
|
| 301 |
+
"learning_rate": 9.783022097502204e-05,
|
| 302 |
+
"loss": 0.0001,
|
| 303 |
+
"step": 49
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.12,
|
| 307 |
+
"learning_rate": 9.772121780783201e-05,
|
| 308 |
+
"loss": 0.0001,
|
| 309 |
+
"step": 50
|
| 310 |
+
},
|
| 311 |
+
{
|
| 312 |
+
"epoch": 0.12,
|
| 313 |
+
"learning_rate": 9.760960709796141e-05,
|
| 314 |
+
"loss": 0.0001,
|
| 315 |
+
"step": 51
|
| 316 |
+
},
|
| 317 |
+
{
|
| 318 |
+
"epoch": 0.12,
|
| 319 |
+
"learning_rate": 9.749539494394886e-05,
|
| 320 |
+
"loss": 0.0001,
|
| 321 |
+
"step": 52
|
| 322 |
+
},
|
| 323 |
+
{
|
| 324 |
+
"epoch": 0.12,
|
| 325 |
+
"learning_rate": 9.737858758647889e-05,
|
| 326 |
+
"loss": 0.0001,
|
| 327 |
+
"step": 53
|
| 328 |
+
},
|
| 329 |
+
{
|
| 330 |
+
"epoch": 0.12,
|
| 331 |
+
"learning_rate": 9.725919140804099e-05,
|
| 332 |
+
"loss": 0.0,
|
| 333 |
+
"step": 54
|
| 334 |
+
},
|
| 335 |
+
{
|
| 336 |
+
"epoch": 0.13,
|
| 337 |
+
"learning_rate": 9.713721293258078e-05,
|
| 338 |
+
"loss": 0.0,
|
| 339 |
+
"step": 55
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 0.13,
|
| 343 |
+
"learning_rate": 9.701265882514361e-05,
|
| 344 |
+
"loss": 0.0,
|
| 345 |
+
"step": 56
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.13,
|
| 349 |
+
"learning_rate": 9.688553589151037e-05,
|
| 350 |
+
"loss": 0.0,
|
| 351 |
+
"step": 57
|
| 352 |
+
},
|
| 353 |
+
{
|
| 354 |
+
"epoch": 0.13,
|
| 355 |
+
"learning_rate": 9.675585107782557e-05,
|
| 356 |
+
"loss": 0.0001,
|
| 357 |
+
"step": 58
|
| 358 |
+
},
|
| 359 |
+
{
|
| 360 |
+
"epoch": 0.14,
|
| 361 |
+
"learning_rate": 9.662361147021779e-05,
|
| 362 |
+
"loss": 0.0,
|
| 363 |
+
"step": 59
|
| 364 |
+
},
|
| 365 |
+
{
|
| 366 |
+
"epoch": 0.14,
|
| 367 |
+
"learning_rate": 9.648882429441257e-05,
|
| 368 |
+
"loss": 0.0,
|
| 369 |
+
"step": 60
|
| 370 |
+
},
|
| 371 |
+
{
|
| 372 |
+
"epoch": 0.14,
|
| 373 |
+
"learning_rate": 9.635149691533749e-05,
|
| 374 |
+
"loss": 0.0,
|
| 375 |
+
"step": 61
|
| 376 |
+
},
|
| 377 |
+
{
|
| 378 |
+
"epoch": 0.14,
|
| 379 |
+
"learning_rate": 9.62116368367198e-05,
|
| 380 |
+
"loss": 0.0,
|
| 381 |
+
"step": 62
|
| 382 |
+
},
|
| 383 |
+
{
|
| 384 |
+
"epoch": 0.15,
|
| 385 |
+
"learning_rate": 9.606925170067638e-05,
|
| 386 |
+
"loss": 0.0,
|
| 387 |
+
"step": 63
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.15,
|
| 391 |
+
"learning_rate": 9.592434928729616e-05,
|
| 392 |
+
"loss": 0.0,
|
| 393 |
+
"step": 64
|
| 394 |
+
},
|
| 395 |
+
{
|
| 396 |
+
"epoch": 0.15,
|
| 397 |
+
"learning_rate": 9.577693751421506e-05,
|
| 398 |
+
"loss": 0.0,
|
| 399 |
+
"step": 65
|
| 400 |
+
},
|
| 401 |
+
{
|
| 402 |
+
"epoch": 0.15,
|
| 403 |
+
"learning_rate": 9.562702443618331e-05,
|
| 404 |
+
"loss": 0.0,
|
| 405 |
+
"step": 66
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 0.15,
|
| 409 |
+
"learning_rate": 9.547461824462534e-05,
|
| 410 |
+
"loss": 0.0,
|
| 411 |
+
"step": 67
|
| 412 |
+
},
|
| 413 |
+
{
|
| 414 |
+
"epoch": 0.16,
|
| 415 |
+
"learning_rate": 9.531972726719216e-05,
|
| 416 |
+
"loss": 0.0,
|
| 417 |
+
"step": 68
|
| 418 |
+
},
|
| 419 |
+
{
|
| 420 |
+
"epoch": 0.16,
|
| 421 |
+
"learning_rate": 9.516235996730645e-05,
|
| 422 |
+
"loss": 0.0,
|
| 423 |
+
"step": 69
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 0.16,
|
| 427 |
+
"learning_rate": 9.500252494369993e-05,
|
| 428 |
+
"loss": 0.0,
|
| 429 |
+
"step": 70
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.16,
|
| 433 |
+
"learning_rate": 9.484023092994366e-05,
|
| 434 |
+
"loss": 0.0,
|
| 435 |
+
"step": 71
|
| 436 |
+
},
|
| 437 |
+
{
|
| 438 |
+
"epoch": 0.17,
|
| 439 |
+
"learning_rate": 9.467548679397072e-05,
|
| 440 |
+
"loss": 0.0,
|
| 441 |
+
"step": 72
|
| 442 |
+
},
|
| 443 |
+
{
|
| 444 |
+
"epoch": 0.17,
|
| 445 |
+
"learning_rate": 9.450830153759177e-05,
|
| 446 |
+
"loss": 0.0,
|
| 447 |
+
"step": 73
|
| 448 |
+
},
|
| 449 |
+
{
|
| 450 |
+
"epoch": 0.17,
|
| 451 |
+
"learning_rate": 9.43386842960031e-05,
|
| 452 |
+
"loss": 0.0,
|
| 453 |
+
"step": 74
|
| 454 |
+
},
|
| 455 |
+
{
|
| 456 |
+
"epoch": 0.17,
|
| 457 |
+
"learning_rate": 9.416664433728748e-05,
|
| 458 |
+
"loss": 0.0,
|
| 459 |
+
"step": 75
|
| 460 |
+
},
|
| 461 |
+
{
|
| 462 |
+
"epoch": 0.18,
|
| 463 |
+
"learning_rate": 9.399219106190775e-05,
|
| 464 |
+
"loss": 0.0,
|
| 465 |
+
"step": 76
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 0.18,
|
| 469 |
+
"learning_rate": 9.381533400219318e-05,
|
| 470 |
+
"loss": 0.0,
|
| 471 |
+
"step": 77
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.18,
|
| 475 |
+
"learning_rate": 9.363608282181861e-05,
|
| 476 |
+
"loss": 0.0,
|
| 477 |
+
"step": 78
|
| 478 |
+
},
|
| 479 |
+
{
|
| 480 |
+
"epoch": 0.18,
|
| 481 |
+
"learning_rate": 9.345444731527642e-05,
|
| 482 |
+
"loss": 0.0,
|
| 483 |
+
"step": 79
|
| 484 |
+
},
|
| 485 |
+
{
|
| 486 |
+
"epoch": 0.18,
|
| 487 |
+
"learning_rate": 9.32704374073413e-05,
|
| 488 |
+
"loss": 0.0,
|
| 489 |
+
"step": 80
|
| 490 |
+
},
|
| 491 |
+
{
|
| 492 |
+
"epoch": 0.19,
|
| 493 |
+
"learning_rate": 9.308406315252799e-05,
|
| 494 |
+
"loss": 0.0172,
|
| 495 |
+
"step": 81
|
| 496 |
+
},
|
| 497 |
+
{
|
| 498 |
+
"epoch": 0.19,
|
| 499 |
+
"learning_rate": 9.289533473454192e-05,
|
| 500 |
+
"loss": 0.0,
|
| 501 |
+
"step": 82
|
| 502 |
+
},
|
| 503 |
+
{
|
| 504 |
+
"epoch": 0.19,
|
| 505 |
+
"learning_rate": 9.270426246572272e-05,
|
| 506 |
+
"loss": 0.0359,
|
| 507 |
+
"step": 83
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 0.19,
|
| 511 |
+
"learning_rate": 9.251085678648072e-05,
|
| 512 |
+
"loss": 0.0,
|
| 513 |
+
"step": 84
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.2,
|
| 517 |
+
"learning_rate": 9.23151282647265e-05,
|
| 518 |
+
"loss": 0.0,
|
| 519 |
+
"step": 85
|
| 520 |
+
},
|
| 521 |
+
{
|
| 522 |
+
"epoch": 0.2,
|
| 523 |
+
"learning_rate": 9.211708759529347e-05,
|
| 524 |
+
"loss": 0.0,
|
| 525 |
+
"step": 86
|
| 526 |
+
},
|
| 527 |
+
{
|
| 528 |
+
"epoch": 0.2,
|
| 529 |
+
"learning_rate": 9.191674559935348e-05,
|
| 530 |
+
"loss": 0.0001,
|
| 531 |
+
"step": 87
|
| 532 |
+
},
|
| 533 |
+
{
|
| 534 |
+
"epoch": 0.2,
|
| 535 |
+
"learning_rate": 9.171411322382552e-05,
|
| 536 |
+
"loss": 0.0,
|
| 537 |
+
"step": 88
|
| 538 |
+
},
|
| 539 |
+
{
|
| 540 |
+
"epoch": 0.21,
|
| 541 |
+
"learning_rate": 9.150920154077754e-05,
|
| 542 |
+
"loss": 0.0,
|
| 543 |
+
"step": 89
|
| 544 |
+
},
|
| 545 |
+
{
|
| 546 |
+
"epoch": 0.21,
|
| 547 |
+
"learning_rate": 9.130202174682155e-05,
|
| 548 |
+
"loss": 0.0001,
|
| 549 |
+
"step": 90
|
| 550 |
+
},
|
| 551 |
+
{
|
| 552 |
+
"epoch": 0.21,
|
| 553 |
+
"learning_rate": 9.109258516250172e-05,
|
| 554 |
+
"loss": 0.0001,
|
| 555 |
+
"step": 91
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.21,
|
| 559 |
+
"learning_rate": 9.08809032316759e-05,
|
| 560 |
+
"loss": 0.0001,
|
| 561 |
+
"step": 92
|
| 562 |
+
},
|
| 563 |
+
{
|
| 564 |
+
"epoch": 0.21,
|
| 565 |
+
"learning_rate": 9.066698752089029e-05,
|
| 566 |
+
"loss": 0.0001,
|
| 567 |
+
"step": 93
|
| 568 |
+
},
|
| 569 |
+
{
|
| 570 |
+
"epoch": 0.22,
|
| 571 |
+
"learning_rate": 9.045084971874738e-05,
|
| 572 |
+
"loss": 0.0001,
|
| 573 |
+
"step": 94
|
| 574 |
+
},
|
| 575 |
+
{
|
| 576 |
+
"epoch": 0.22,
|
| 577 |
+
"learning_rate": 9.023250163526731e-05,
|
| 578 |
+
"loss": 0.0001,
|
| 579 |
+
"step": 95
|
| 580 |
+
},
|
| 581 |
+
{
|
| 582 |
+
"epoch": 0.22,
|
| 583 |
+
"learning_rate": 9.001195520124255e-05,
|
| 584 |
+
"loss": 0.0001,
|
| 585 |
+
"step": 96
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 0.22,
|
| 589 |
+
"learning_rate": 8.978922246758606e-05,
|
| 590 |
+
"loss": 0.1254,
|
| 591 |
+
"step": 97
|
| 592 |
+
},
|
| 593 |
+
{
|
| 594 |
+
"epoch": 0.23,
|
| 595 |
+
"learning_rate": 8.956431560467267e-05,
|
| 596 |
+
"loss": 0.0001,
|
| 597 |
+
"step": 98
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 0.23,
|
| 601 |
+
"learning_rate": 8.933724690167417e-05,
|
| 602 |
+
"loss": 0.1316,
|
| 603 |
+
"step": 99
|
| 604 |
+
},
|
| 605 |
+
{
|
| 606 |
+
"epoch": 0.23,
|
| 607 |
+
"learning_rate": 8.910802876588781e-05,
|
| 608 |
+
"loss": 0.0001,
|
| 609 |
+
"step": 100
|
| 610 |
+
},
|
| 611 |
+
{
|
| 612 |
+
"epoch": 0.23,
|
| 613 |
+
"learning_rate": 8.887667372205839e-05,
|
| 614 |
+
"loss": 0.0001,
|
| 615 |
+
"step": 101
|
| 616 |
+
},
|
| 617 |
+
{
|
| 618 |
+
"epoch": 0.24,
|
| 619 |
+
"learning_rate": 8.864319441169373e-05,
|
| 620 |
+
"loss": 0.0001,
|
| 621 |
+
"step": 102
|
| 622 |
+
},
|
| 623 |
+
{
|
| 624 |
+
"epoch": 0.24,
|
| 625 |
+
"learning_rate": 8.840760359237412e-05,
|
| 626 |
+
"loss": 0.0002,
|
| 627 |
+
"step": 103
|
| 628 |
+
},
|
| 629 |
+
{
|
| 630 |
+
"epoch": 0.24,
|
| 631 |
+
"learning_rate": 8.816991413705516e-05,
|
| 632 |
+
"loss": 0.0002,
|
| 633 |
+
"step": 104
|
| 634 |
+
},
|
| 635 |
+
{
|
| 636 |
+
"epoch": 0.24,
|
| 637 |
+
"learning_rate": 8.793013903336428e-05,
|
| 638 |
+
"loss": 0.0002,
|
| 639 |
+
"step": 105
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.24,
|
| 643 |
+
"learning_rate": 8.768829138289124e-05,
|
| 644 |
+
"loss": 0.0002,
|
| 645 |
+
"step": 106
|
| 646 |
+
},
|
| 647 |
+
{
|
| 648 |
+
"epoch": 0.25,
|
| 649 |
+
"learning_rate": 8.744438440047206e-05,
|
| 650 |
+
"loss": 0.0002,
|
| 651 |
+
"step": 107
|
| 652 |
+
},
|
| 653 |
+
{
|
| 654 |
+
"epoch": 0.25,
|
| 655 |
+
"learning_rate": 8.719843141346718e-05,
|
| 656 |
+
"loss": 0.0001,
|
| 657 |
+
"step": 108
|
| 658 |
+
},
|
| 659 |
+
{
|
| 660 |
+
"epoch": 0.25,
|
| 661 |
+
"learning_rate": 8.695044586103296e-05,
|
| 662 |
+
"loss": 0.0002,
|
| 663 |
+
"step": 109
|
| 664 |
+
},
|
| 665 |
+
{
|
| 666 |
+
"epoch": 0.25,
|
| 667 |
+
"learning_rate": 8.670044129338761e-05,
|
| 668 |
+
"loss": 0.0001,
|
| 669 |
+
"step": 110
|
| 670 |
+
},
|
| 671 |
+
{
|
| 672 |
+
"epoch": 0.26,
|
| 673 |
+
"learning_rate": 8.644843137107059e-05,
|
| 674 |
+
"loss": 0.0001,
|
| 675 |
+
"step": 111
|
| 676 |
+
},
|
| 677 |
+
{
|
| 678 |
+
"epoch": 0.26,
|
| 679 |
+
"learning_rate": 8.61944298641963e-05,
|
| 680 |
+
"loss": 0.0001,
|
| 681 |
+
"step": 112
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 0.26,
|
| 685 |
+
"learning_rate": 8.593845065170164e-05,
|
| 686 |
+
"loss": 0.0001,
|
| 687 |
+
"step": 113
|
| 688 |
+
},
|
| 689 |
+
{
|
| 690 |
+
"epoch": 0.26,
|
| 691 |
+
"learning_rate": 8.568050772058762e-05,
|
| 692 |
+
"loss": 0.0001,
|
| 693 |
+
"step": 114
|
| 694 |
+
},
|
| 695 |
+
{
|
| 696 |
+
"epoch": 0.26,
|
| 697 |
+
"learning_rate": 8.542061516515512e-05,
|
| 698 |
+
"loss": 0.0001,
|
| 699 |
+
"step": 115
|
| 700 |
+
},
|
| 701 |
+
{
|
| 702 |
+
"epoch": 0.27,
|
| 703 |
+
"learning_rate": 8.515878718623473e-05,
|
| 704 |
+
"loss": 0.0,
|
| 705 |
+
"step": 116
|
| 706 |
+
},
|
| 707 |
+
{
|
| 708 |
+
"epoch": 0.27,
|
| 709 |
+
"learning_rate": 8.489503809041088e-05,
|
| 710 |
+
"loss": 0.1113,
|
| 711 |
+
"step": 117
|
| 712 |
+
},
|
| 713 |
+
{
|
| 714 |
+
"epoch": 0.27,
|
| 715 |
+
"learning_rate": 8.462938228924e-05,
|
| 716 |
+
"loss": 0.0,
|
| 717 |
+
"step": 118
|
| 718 |
+
},
|
| 719 |
+
{
|
| 720 |
+
"epoch": 0.27,
|
| 721 |
+
"learning_rate": 8.436183429846313e-05,
|
| 722 |
+
"loss": 0.0,
|
| 723 |
+
"step": 119
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 0.28,
|
| 727 |
+
"learning_rate": 8.409240873721277e-05,
|
| 728 |
+
"loss": 0.0,
|
| 729 |
+
"step": 120
|
| 730 |
+
},
|
| 731 |
+
{
|
| 732 |
+
"epoch": 0.28,
|
| 733 |
+
"learning_rate": 8.382112032721398e-05,
|
| 734 |
+
"loss": 0.0,
|
| 735 |
+
"step": 121
|
| 736 |
+
},
|
| 737 |
+
{
|
| 738 |
+
"epoch": 0.28,
|
| 739 |
+
"learning_rate": 8.354798389198013e-05,
|
| 740 |
+
"loss": 0.0,
|
| 741 |
+
"step": 122
|
| 742 |
+
},
|
| 743 |
+
{
|
| 744 |
+
"epoch": 0.28,
|
| 745 |
+
"learning_rate": 8.327301435600273e-05,
|
| 746 |
+
"loss": 0.0,
|
| 747 |
+
"step": 123
|
| 748 |
+
},
|
| 749 |
+
{
|
| 750 |
+
"epoch": 0.29,
|
| 751 |
+
"learning_rate": 8.299622674393614e-05,
|
| 752 |
+
"loss": 0.0893,
|
| 753 |
+
"step": 124
|
| 754 |
+
},
|
| 755 |
+
{
|
| 756 |
+
"epoch": 0.29,
|
| 757 |
+
"learning_rate": 8.271763617977641e-05,
|
| 758 |
+
"loss": 0.0,
|
| 759 |
+
"step": 125
|
| 760 |
+
},
|
| 761 |
+
{
|
| 762 |
+
"epoch": 0.29,
|
| 763 |
+
"learning_rate": 8.243725788603509e-05,
|
| 764 |
+
"loss": 0.0,
|
| 765 |
+
"step": 126
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 0.29,
|
| 769 |
+
"learning_rate": 8.215510718290725e-05,
|
| 770 |
+
"loss": 0.0,
|
| 771 |
+
"step": 127
|
| 772 |
+
},
|
| 773 |
+
{
|
| 774 |
+
"epoch": 0.29,
|
| 775 |
+
"learning_rate": 8.18711994874345e-05,
|
| 776 |
+
"loss": 0.0,
|
| 777 |
+
"step": 128
|
| 778 |
+
},
|
| 779 |
+
{
|
| 780 |
+
"epoch": 0.3,
|
| 781 |
+
"learning_rate": 8.158555031266254e-05,
|
| 782 |
+
"loss": 0.0,
|
| 783 |
+
"step": 129
|
| 784 |
+
},
|
| 785 |
+
{
|
| 786 |
+
"epoch": 0.3,
|
| 787 |
+
"learning_rate": 8.129817526679358e-05,
|
| 788 |
+
"loss": 0.0,
|
| 789 |
+
"step": 130
|
| 790 |
+
},
|
| 791 |
+
{
|
| 792 |
+
"epoch": 0.3,
|
| 793 |
+
"learning_rate": 8.100909005233334e-05,
|
| 794 |
+
"loss": 0.0,
|
| 795 |
+
"step": 131
|
| 796 |
+
},
|
| 797 |
+
{
|
| 798 |
+
"epoch": 0.3,
|
| 799 |
+
"learning_rate": 8.071831046523319e-05,
|
| 800 |
+
"loss": 0.0,
|
| 801 |
+
"step": 132
|
| 802 |
+
},
|
| 803 |
+
{
|
| 804 |
+
"epoch": 0.31,
|
| 805 |
+
"learning_rate": 8.042585239402697e-05,
|
| 806 |
+
"loss": 0.0752,
|
| 807 |
+
"step": 133
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 0.31,
|
| 811 |
+
"learning_rate": 8.013173181896283e-05,
|
| 812 |
+
"loss": 0.0,
|
| 813 |
+
"step": 134
|
| 814 |
+
},
|
| 815 |
+
{
|
| 816 |
+
"epoch": 0.31,
|
| 817 |
+
"learning_rate": 7.983596481113005e-05,
|
| 818 |
+
"loss": 0.0,
|
| 819 |
+
"step": 135
|
| 820 |
+
},
|
| 821 |
+
{
|
| 822 |
+
"epoch": 0.31,
|
| 823 |
+
"learning_rate": 7.953856753158096e-05,
|
| 824 |
+
"loss": 0.0,
|
| 825 |
+
"step": 136
|
| 826 |
+
},
|
| 827 |
+
{
|
| 828 |
+
"epoch": 0.32,
|
| 829 |
+
"learning_rate": 7.923955623044775e-05,
|
| 830 |
+
"loss": 0.0,
|
| 831 |
+
"step": 137
|
| 832 |
+
},
|
| 833 |
+
{
|
| 834 |
+
"epoch": 0.32,
|
| 835 |
+
"learning_rate": 7.893894724605469e-05,
|
| 836 |
+
"loss": 0.0,
|
| 837 |
+
"step": 138
|
| 838 |
+
},
|
| 839 |
+
{
|
| 840 |
+
"epoch": 0.32,
|
| 841 |
+
"learning_rate": 7.863675700402526e-05,
|
| 842 |
+
"loss": 0.0,
|
| 843 |
+
"step": 139
|
| 844 |
+
},
|
| 845 |
+
{
|
| 846 |
+
"epoch": 0.32,
|
| 847 |
+
"learning_rate": 7.833300201638475e-05,
|
| 848 |
+
"loss": 0.0,
|
| 849 |
+
"step": 140
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 0.32,
|
| 853 |
+
"learning_rate": 7.80276988806579e-05,
|
| 854 |
+
"loss": 0.0,
|
| 855 |
+
"step": 141
|
| 856 |
+
},
|
| 857 |
+
{
|
| 858 |
+
"epoch": 0.33,
|
| 859 |
+
"learning_rate": 7.772086427896211e-05,
|
| 860 |
+
"loss": 0.0,
|
| 861 |
+
"step": 142
|
| 862 |
+
},
|
| 863 |
+
{
|
| 864 |
+
"epoch": 0.33,
|
| 865 |
+
"learning_rate": 7.741251497709583e-05,
|
| 866 |
+
"loss": 0.0,
|
| 867 |
+
"step": 143
|
| 868 |
+
},
|
| 869 |
+
{
|
| 870 |
+
"epoch": 0.33,
|
| 871 |
+
"learning_rate": 7.710266782362247e-05,
|
| 872 |
+
"loss": 0.0,
|
| 873 |
+
"step": 144
|
| 874 |
+
},
|
| 875 |
+
{
|
| 876 |
+
"epoch": 0.33,
|
| 877 |
+
"learning_rate": 7.679133974894983e-05,
|
| 878 |
+
"loss": 0.0,
|
| 879 |
+
"step": 145
|
| 880 |
+
},
|
| 881 |
+
{
|
| 882 |
+
"epoch": 0.34,
|
| 883 |
+
"learning_rate": 7.647854776440496e-05,
|
| 884 |
+
"loss": 0.0,
|
| 885 |
+
"step": 146
|
| 886 |
+
},
|
| 887 |
+
{
|
| 888 |
+
"epoch": 0.34,
|
| 889 |
+
"learning_rate": 7.616430896130455e-05,
|
| 890 |
+
"loss": 0.0,
|
| 891 |
+
"step": 147
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 0.34,
|
| 895 |
+
"learning_rate": 7.584864051002128e-05,
|
| 896 |
+
"loss": 0.0,
|
| 897 |
+
"step": 148
|
| 898 |
+
},
|
| 899 |
+
{
|
| 900 |
+
"epoch": 0.34,
|
| 901 |
+
"learning_rate": 7.553155965904535e-05,
|
| 902 |
+
"loss": 0.0,
|
| 903 |
+
"step": 149
|
| 904 |
+
},
|
| 905 |
+
{
|
| 906 |
+
"epoch": 0.35,
|
| 907 |
+
"learning_rate": 7.521308373404217e-05,
|
| 908 |
+
"loss": 0.0,
|
| 909 |
+
"step": 150
|
| 910 |
+
},
|
| 911 |
+
{
|
| 912 |
+
"epoch": 0.35,
|
| 913 |
+
"learning_rate": 7.489323013690562e-05,
|
| 914 |
+
"loss": 0.0,
|
| 915 |
+
"step": 151
|
| 916 |
+
},
|
| 917 |
+
{
|
| 918 |
+
"epoch": 0.35,
|
| 919 |
+
"learning_rate": 7.457201634480713e-05,
|
| 920 |
+
"loss": 0.0,
|
| 921 |
+
"step": 152
|
| 922 |
+
},
|
| 923 |
+
{
|
| 924 |
+
"epoch": 0.35,
|
| 925 |
+
"learning_rate": 7.424945990924079e-05,
|
| 926 |
+
"loss": 0.0,
|
| 927 |
+
"step": 153
|
| 928 |
+
},
|
| 929 |
+
{
|
| 930 |
+
"epoch": 0.35,
|
| 931 |
+
"learning_rate": 7.392557845506432e-05,
|
| 932 |
+
"loss": 0.0,
|
| 933 |
+
"step": 154
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 0.36,
|
| 937 |
+
"learning_rate": 7.360038967953598e-05,
|
| 938 |
+
"loss": 0.0,
|
| 939 |
+
"step": 155
|
| 940 |
+
},
|
| 941 |
+
{
|
| 942 |
+
"epoch": 0.36,
|
| 943 |
+
"learning_rate": 7.327391135134748e-05,
|
| 944 |
+
"loss": 0.0,
|
| 945 |
+
"step": 156
|
| 946 |
+
},
|
| 947 |
+
{
|
| 948 |
+
"epoch": 0.36,
|
| 949 |
+
"learning_rate": 7.294616130965337e-05,
|
| 950 |
+
"loss": 0.0,
|
| 951 |
+
"step": 157
|
| 952 |
+
},
|
| 953 |
+
{
|
| 954 |
+
"epoch": 0.36,
|
| 955 |
+
"learning_rate": 7.261715746309595e-05,
|
| 956 |
+
"loss": 0.0,
|
| 957 |
+
"step": 158
|
| 958 |
+
},
|
| 959 |
+
{
|
| 960 |
+
"epoch": 0.37,
|
| 961 |
+
"learning_rate": 7.228691778882693e-05,
|
| 962 |
+
"loss": 0.0,
|
| 963 |
+
"step": 159
|
| 964 |
+
},
|
| 965 |
+
{
|
| 966 |
+
"epoch": 0.37,
|
| 967 |
+
"learning_rate": 7.195546033152505e-05,
|
| 968 |
+
"loss": 0.0,
|
| 969 |
+
"step": 160
|
| 970 |
+
},
|
| 971 |
+
{
|
| 972 |
+
"epoch": 0.37,
|
| 973 |
+
"learning_rate": 7.16228032024102e-05,
|
| 974 |
+
"loss": 0.0,
|
| 975 |
+
"step": 161
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 0.37,
|
| 979 |
+
"learning_rate": 7.128896457825364e-05,
|
| 980 |
+
"loss": 0.0,
|
| 981 |
+
"step": 162
|
| 982 |
+
},
|
| 983 |
+
{
|
| 984 |
+
"epoch": 0.38,
|
| 985 |
+
"learning_rate": 7.095396270038493e-05,
|
| 986 |
+
"loss": 0.0,
|
| 987 |
+
"step": 163
|
| 988 |
+
},
|
| 989 |
+
{
|
| 990 |
+
"epoch": 0.38,
|
| 991 |
+
"learning_rate": 7.061781587369519e-05,
|
| 992 |
+
"loss": 0.0,
|
| 993 |
+
"step": 164
|
| 994 |
+
},
|
| 995 |
+
{
|
| 996 |
+
"epoch": 0.38,
|
| 997 |
+
"learning_rate": 7.02805424656368e-05,
|
| 998 |
+
"loss": 0.0,
|
| 999 |
+
"step": 165
|
| 1000 |
+
},
|
| 1001 |
+
{
|
| 1002 |
+
"epoch": 0.38,
|
| 1003 |
+
"learning_rate": 6.994216090521991e-05,
|
| 1004 |
+
"loss": 0.0,
|
| 1005 |
+
"step": 166
|
| 1006 |
+
},
|
| 1007 |
+
{
|
| 1008 |
+
"epoch": 0.38,
|
| 1009 |
+
"learning_rate": 6.960268968200538e-05,
|
| 1010 |
+
"loss": 0.0,
|
| 1011 |
+
"step": 167
|
| 1012 |
+
},
|
| 1013 |
+
{
|
| 1014 |
+
"epoch": 0.39,
|
| 1015 |
+
"learning_rate": 6.92621473450945e-05,
|
| 1016 |
+
"loss": 0.0,
|
| 1017 |
+
"step": 168
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 0.39,
|
| 1021 |
+
"learning_rate": 6.892055250211552e-05,
|
| 1022 |
+
"loss": 0.0,
|
| 1023 |
+
"step": 169
|
| 1024 |
+
},
|
| 1025 |
+
{
|
| 1026 |
+
"epoch": 0.39,
|
| 1027 |
+
"learning_rate": 6.857792381820672e-05,
|
| 1028 |
+
"loss": 0.0,
|
| 1029 |
+
"step": 170
|
| 1030 |
+
},
|
| 1031 |
+
{
|
| 1032 |
+
"epoch": 0.39,
|
| 1033 |
+
"learning_rate": 6.823428001499677e-05,
|
| 1034 |
+
"loss": 0.0,
|
| 1035 |
+
"step": 171
|
| 1036 |
+
},
|
| 1037 |
+
{
|
| 1038 |
+
"epoch": 0.4,
|
| 1039 |
+
"learning_rate": 6.788963986958153e-05,
|
| 1040 |
+
"loss": 0.0,
|
| 1041 |
+
"step": 172
|
| 1042 |
+
},
|
| 1043 |
+
{
|
| 1044 |
+
"epoch": 0.4,
|
| 1045 |
+
"learning_rate": 6.754402221349824e-05,
|
| 1046 |
+
"loss": 0.0,
|
| 1047 |
+
"step": 173
|
| 1048 |
+
},
|
| 1049 |
+
{
|
| 1050 |
+
"epoch": 0.4,
|
| 1051 |
+
"learning_rate": 6.719744593169641e-05,
|
| 1052 |
+
"loss": 0.0,
|
| 1053 |
+
"step": 174
|
| 1054 |
+
},
|
| 1055 |
+
{
|
| 1056 |
+
"epoch": 0.4,
|
| 1057 |
+
"learning_rate": 6.684992996150599e-05,
|
| 1058 |
+
"loss": 0.0,
|
| 1059 |
+
"step": 175
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 0.41,
|
| 1063 |
+
"learning_rate": 6.650149329160259e-05,
|
| 1064 |
+
"loss": 0.0,
|
| 1065 |
+
"step": 176
|
| 1066 |
+
},
|
| 1067 |
+
{
|
| 1068 |
+
"epoch": 0.41,
|
| 1069 |
+
"learning_rate": 6.615215496096986e-05,
|
| 1070 |
+
"loss": 0.0,
|
| 1071 |
+
"step": 177
|
| 1072 |
+
},
|
| 1073 |
+
{
|
| 1074 |
+
"epoch": 0.41,
|
| 1075 |
+
"learning_rate": 6.580193405785938e-05,
|
| 1076 |
+
"loss": 0.0,
|
| 1077 |
+
"step": 178
|
| 1078 |
+
},
|
| 1079 |
+
{
|
| 1080 |
+
"epoch": 0.41,
|
| 1081 |
+
"learning_rate": 6.545084971874738e-05,
|
| 1082 |
+
"loss": 0.0,
|
| 1083 |
+
"step": 179
|
| 1084 |
+
},
|
| 1085 |
+
{
|
| 1086 |
+
"epoch": 0.41,
|
| 1087 |
+
"learning_rate": 6.509892112728928e-05,
|
| 1088 |
+
"loss": 0.0,
|
| 1089 |
+
"step": 180
|
| 1090 |
+
},
|
| 1091 |
+
{
|
| 1092 |
+
"epoch": 0.42,
|
| 1093 |
+
"learning_rate": 6.474616751327142e-05,
|
| 1094 |
+
"loss": 0.0,
|
| 1095 |
+
"step": 181
|
| 1096 |
+
},
|
| 1097 |
+
{
|
| 1098 |
+
"epoch": 0.42,
|
| 1099 |
+
"learning_rate": 6.439260815156039e-05,
|
| 1100 |
+
"loss": 0.0,
|
| 1101 |
+
"step": 182
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 0.42,
|
| 1105 |
+
"learning_rate": 6.403826236104965e-05,
|
| 1106 |
+
"loss": 0.0,
|
| 1107 |
+
"step": 183
|
| 1108 |
+
},
|
| 1109 |
+
{
|
| 1110 |
+
"epoch": 0.42,
|
| 1111 |
+
"learning_rate": 6.368314950360415e-05,
|
| 1112 |
+
"loss": 0.0,
|
| 1113 |
+
"step": 184
|
| 1114 |
+
},
|
| 1115 |
+
{
|
| 1116 |
+
"epoch": 0.43,
|
| 1117 |
+
"learning_rate": 6.332728898300221e-05,
|
| 1118 |
+
"loss": 0.0,
|
| 1119 |
+
"step": 185
|
| 1120 |
+
},
|
| 1121 |
+
{
|
| 1122 |
+
"epoch": 0.43,
|
| 1123 |
+
"learning_rate": 6.297070024387535e-05,
|
| 1124 |
+
"loss": 0.0,
|
| 1125 |
+
"step": 186
|
| 1126 |
+
},
|
| 1127 |
+
{
|
| 1128 |
+
"epoch": 0.43,
|
| 1129 |
+
"learning_rate": 6.261340277064578e-05,
|
| 1130 |
+
"loss": 0.0,
|
| 1131 |
+
"step": 187
|
| 1132 |
+
},
|
| 1133 |
+
{
|
| 1134 |
+
"epoch": 0.43,
|
| 1135 |
+
"learning_rate": 6.225541608646179e-05,
|
| 1136 |
+
"loss": 0.0,
|
| 1137 |
+
"step": 188
|
| 1138 |
+
},
|
| 1139 |
+
{
|
| 1140 |
+
"epoch": 0.44,
|
| 1141 |
+
"learning_rate": 6.189675975213094e-05,
|
| 1142 |
+
"loss": 0.0,
|
| 1143 |
+
"step": 189
|
| 1144 |
+
},
|
| 1145 |
+
{
|
| 1146 |
+
"epoch": 0.44,
|
| 1147 |
+
"learning_rate": 6.153745336505125e-05,
|
| 1148 |
+
"loss": 0.0,
|
| 1149 |
+
"step": 190
|
| 1150 |
+
},
|
| 1151 |
+
{
|
| 1152 |
+
"epoch": 0.44,
|
| 1153 |
+
"learning_rate": 6.117751655814037e-05,
|
| 1154 |
+
"loss": 0.0,
|
| 1155 |
+
"step": 191
|
| 1156 |
+
},
|
| 1157 |
+
{
|
| 1158 |
+
"epoch": 0.44,
|
| 1159 |
+
"learning_rate": 6.081696899876282e-05,
|
| 1160 |
+
"loss": 0.0,
|
| 1161 |
+
"step": 192
|
| 1162 |
+
},
|
| 1163 |
+
{
|
| 1164 |
+
"epoch": 0.44,
|
| 1165 |
+
"learning_rate": 6.0455830387655376e-05,
|
| 1166 |
+
"loss": 0.0,
|
| 1167 |
+
"step": 193
|
| 1168 |
+
},
|
| 1169 |
+
{
|
| 1170 |
+
"epoch": 0.45,
|
| 1171 |
+
"learning_rate": 6.009412045785051e-05,
|
| 1172 |
+
"loss": 0.0,
|
| 1173 |
+
"step": 194
|
| 1174 |
+
},
|
| 1175 |
+
{
|
| 1176 |
+
"epoch": 0.45,
|
| 1177 |
+
"learning_rate": 5.973185897359828e-05,
|
| 1178 |
+
"loss": 0.0,
|
| 1179 |
+
"step": 195
|
| 1180 |
+
},
|
| 1181 |
+
{
|
| 1182 |
+
"epoch": 0.45,
|
| 1183 |
+
"learning_rate": 5.9369065729286245e-05,
|
| 1184 |
+
"loss": 0.0,
|
| 1185 |
+
"step": 196
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 0.45,
|
| 1189 |
+
"learning_rate": 5.900576054835798e-05,
|
| 1190 |
+
"loss": 0.0,
|
| 1191 |
+
"step": 197
|
| 1192 |
+
},
|
| 1193 |
+
{
|
| 1194 |
+
"epoch": 0.46,
|
| 1195 |
+
"learning_rate": 5.8641963282229905e-05,
|
| 1196 |
+
"loss": 0.0,
|
| 1197 |
+
"step": 198
|
| 1198 |
+
},
|
| 1199 |
+
{
|
| 1200 |
+
"epoch": 0.46,
|
| 1201 |
+
"learning_rate": 5.82776938092065e-05,
|
| 1202 |
+
"loss": 0.0,
|
| 1203 |
+
"step": 199
|
| 1204 |
+
},
|
| 1205 |
+
{
|
| 1206 |
+
"epoch": 0.46,
|
| 1207 |
+
"learning_rate": 5.7912972033394195e-05,
|
| 1208 |
+
"loss": 0.0,
|
| 1209 |
+
"step": 200
|
| 1210 |
+
},
|
| 1211 |
+
{
|
| 1212 |
+
"epoch": 0.46,
|
| 1213 |
+
"learning_rate": 5.7547817883613764e-05,
|
| 1214 |
+
"loss": 0.0358,
|
| 1215 |
+
"step": 201
|
| 1216 |
+
},
|
| 1217 |
+
{
|
| 1218 |
+
"epoch": 0.47,
|
| 1219 |
+
"learning_rate": 5.718225131231142e-05,
|
| 1220 |
+
"loss": 0.0,
|
| 1221 |
+
"step": 202
|
| 1222 |
+
},
|
| 1223 |
+
{
|
| 1224 |
+
"epoch": 0.47,
|
| 1225 |
+
"learning_rate": 5.6816292294468495e-05,
|
| 1226 |
+
"loss": 0.0,
|
| 1227 |
+
"step": 203
|
| 1228 |
+
},
|
| 1229 |
+
{
|
| 1230 |
+
"epoch": 0.47,
|
| 1231 |
+
"learning_rate": 5.644996082651017e-05,
|
| 1232 |
+
"loss": 0.0676,
|
| 1233 |
+
"step": 204
|
| 1234 |
+
},
|
| 1235 |
+
{
|
| 1236 |
+
"epoch": 0.47,
|
| 1237 |
+
"learning_rate": 5.6083276925212624e-05,
|
| 1238 |
+
"loss": 0.0,
|
| 1239 |
+
"step": 205
|
| 1240 |
+
},
|
| 1241 |
+
{
|
| 1242 |
+
"epoch": 0.47,
|
| 1243 |
+
"learning_rate": 5.571626062660946e-05,
|
| 1244 |
+
"loss": 0.0,
|
| 1245 |
+
"step": 206
|
| 1246 |
+
},
|
| 1247 |
+
{
|
| 1248 |
+
"epoch": 0.48,
|
| 1249 |
+
"learning_rate": 5.5348931984896836e-05,
|
| 1250 |
+
"loss": 0.0,
|
| 1251 |
+
"step": 207
|
| 1252 |
+
},
|
| 1253 |
+
{
|
| 1254 |
+
"epoch": 0.48,
|
| 1255 |
+
"learning_rate": 5.49813110713377e-05,
|
| 1256 |
+
"loss": 0.0743,
|
| 1257 |
+
"step": 208
|
| 1258 |
+
},
|
| 1259 |
+
{
|
| 1260 |
+
"epoch": 0.48,
|
| 1261 |
+
"learning_rate": 5.4613417973165106e-05,
|
| 1262 |
+
"loss": 0.0,
|
| 1263 |
+
"step": 209
|
| 1264 |
+
},
|
| 1265 |
+
{
|
| 1266 |
+
"epoch": 0.48,
|
| 1267 |
+
"learning_rate": 5.4245272792484524e-05,
|
| 1268 |
+
"loss": 0.0,
|
| 1269 |
+
"step": 210
|
| 1270 |
+
},
|
| 1271 |
+
{
|
| 1272 |
+
"epoch": 0.49,
|
| 1273 |
+
"learning_rate": 5.387689564517558e-05,
|
| 1274 |
+
"loss": 0.0,
|
| 1275 |
+
"step": 211
|
| 1276 |
+
},
|
| 1277 |
+
{
|
| 1278 |
+
"epoch": 0.49,
|
| 1279 |
+
"learning_rate": 5.3508306659792764e-05,
|
| 1280 |
+
"loss": 0.0,
|
| 1281 |
+
"step": 212
|
| 1282 |
+
},
|
| 1283 |
+
{
|
| 1284 |
+
"epoch": 0.49,
|
| 1285 |
+
"learning_rate": 5.313952597646568e-05,
|
| 1286 |
+
"loss": 0.0,
|
| 1287 |
+
"step": 213
|
| 1288 |
+
},
|
| 1289 |
+
{
|
| 1290 |
+
"epoch": 0.49,
|
| 1291 |
+
"learning_rate": 5.27705737457985e-05,
|
| 1292 |
+
"loss": 0.0,
|
| 1293 |
+
"step": 214
|
| 1294 |
+
},
|
| 1295 |
+
{
|
| 1296 |
+
"epoch": 0.5,
|
| 1297 |
+
"learning_rate": 5.240147012776899e-05,
|
| 1298 |
+
"loss": 0.0,
|
| 1299 |
+
"step": 215
|
| 1300 |
+
},
|
| 1301 |
+
{
|
| 1302 |
+
"epoch": 0.5,
|
| 1303 |
+
"learning_rate": 5.203223529062684e-05,
|
| 1304 |
+
"loss": 0.0,
|
| 1305 |
+
"step": 216
|
| 1306 |
+
},
|
| 1307 |
+
{
|
| 1308 |
+
"epoch": 0.5,
|
| 1309 |
+
"learning_rate": 5.166288940979175e-05,
|
| 1310 |
+
"loss": 0.0,
|
| 1311 |
+
"step": 217
|
| 1312 |
+
},
|
| 1313 |
+
{
|
| 1314 |
+
"epoch": 0.5,
|
| 1315 |
+
"learning_rate": 5.129345266675095e-05,
|
| 1316 |
+
"loss": 0.0,
|
| 1317 |
+
"step": 218
|
| 1318 |
+
},
|
| 1319 |
+
{
|
| 1320 |
+
"epoch": 0.5,
|
| 1321 |
+
"learning_rate": 5.092394524795649e-05,
|
| 1322 |
+
"loss": 0.0,
|
| 1323 |
+
"step": 219
|
| 1324 |
+
},
|
| 1325 |
+
{
|
| 1326 |
+
"epoch": 0.51,
|
| 1327 |
+
"learning_rate": 5.055438734372225e-05,
|
| 1328 |
+
"loss": 0.0,
|
| 1329 |
+
"step": 220
|
| 1330 |
+
},
|
| 1331 |
+
{
|
| 1332 |
+
"epoch": 0.51,
|
| 1333 |
+
"learning_rate": 5.0184799147120656e-05,
|
| 1334 |
+
"loss": 0.0,
|
| 1335 |
+
"step": 221
|
| 1336 |
+
},
|
| 1337 |
+
{
|
| 1338 |
+
"epoch": 0.51,
|
| 1339 |
+
"learning_rate": 4.981520085287935e-05,
|
| 1340 |
+
"loss": 0.0,
|
| 1341 |
+
"step": 222
|
| 1342 |
+
},
|
| 1343 |
+
{
|
| 1344 |
+
"epoch": 0.51,
|
| 1345 |
+
"learning_rate": 4.944561265627776e-05,
|
| 1346 |
+
"loss": 0.0625,
|
| 1347 |
+
"step": 223
|
| 1348 |
+
},
|
| 1349 |
+
{
|
| 1350 |
+
"epoch": 0.52,
|
| 1351 |
+
"learning_rate": 4.907605475204352e-05,
|
| 1352 |
+
"loss": 0.0,
|
| 1353 |
+
"step": 224
|
| 1354 |
+
},
|
| 1355 |
+
{
|
| 1356 |
+
"epoch": 0.52,
|
| 1357 |
+
"learning_rate": 4.870654733324907e-05,
|
| 1358 |
+
"loss": 0.0,
|
| 1359 |
+
"step": 225
|
| 1360 |
+
},
|
| 1361 |
+
{
|
| 1362 |
+
"epoch": 0.52,
|
| 1363 |
+
"learning_rate": 4.833711059020826e-05,
|
| 1364 |
+
"loss": 0.0,
|
| 1365 |
+
"step": 226
|
| 1366 |
+
},
|
| 1367 |
+
{
|
| 1368 |
+
"epoch": 0.52,
|
| 1369 |
+
"learning_rate": 4.796776470937317e-05,
|
| 1370 |
+
"loss": 0.0,
|
| 1371 |
+
"step": 227
|
| 1372 |
+
},
|
| 1373 |
+
{
|
| 1374 |
+
"epoch": 0.53,
|
| 1375 |
+
"learning_rate": 4.759852987223103e-05,
|
| 1376 |
+
"loss": 0.0,
|
| 1377 |
+
"step": 228
|
| 1378 |
+
},
|
| 1379 |
+
{
|
| 1380 |
+
"epoch": 0.53,
|
| 1381 |
+
"learning_rate": 4.72294262542015e-05,
|
| 1382 |
+
"loss": 0.0,
|
| 1383 |
+
"step": 229
|
| 1384 |
+
},
|
| 1385 |
+
{
|
| 1386 |
+
"epoch": 0.53,
|
| 1387 |
+
"learning_rate": 4.6860474023534335e-05,
|
| 1388 |
+
"loss": 0.0,
|
| 1389 |
+
"step": 230
|
| 1390 |
+
},
|
| 1391 |
+
{
|
| 1392 |
+
"epoch": 0.53,
|
| 1393 |
+
"learning_rate": 4.6491693340207254e-05,
|
| 1394 |
+
"loss": 0.0,
|
| 1395 |
+
"step": 231
|
| 1396 |
+
},
|
| 1397 |
+
{
|
| 1398 |
+
"epoch": 0.53,
|
| 1399 |
+
"learning_rate": 4.612310435482443e-05,
|
| 1400 |
+
"loss": 0.0,
|
| 1401 |
+
"step": 232
|
| 1402 |
+
},
|
| 1403 |
+
{
|
| 1404 |
+
"epoch": 0.54,
|
| 1405 |
+
"learning_rate": 4.575472720751547e-05,
|
| 1406 |
+
"loss": 0.0,
|
| 1407 |
+
"step": 233
|
| 1408 |
+
},
|
| 1409 |
+
{
|
| 1410 |
+
"epoch": 0.54,
|
| 1411 |
+
"learning_rate": 4.5386582026834906e-05,
|
| 1412 |
+
"loss": 0.0,
|
| 1413 |
+
"step": 234
|
| 1414 |
+
},
|
| 1415 |
+
{
|
| 1416 |
+
"epoch": 0.54,
|
| 1417 |
+
"learning_rate": 4.50186889286623e-05,
|
| 1418 |
+
"loss": 0.0,
|
| 1419 |
+
"step": 235
|
| 1420 |
+
},
|
| 1421 |
+
{
|
| 1422 |
+
"epoch": 0.54,
|
| 1423 |
+
"learning_rate": 4.465106801510319e-05,
|
| 1424 |
+
"loss": 0.0,
|
| 1425 |
+
"step": 236
|
| 1426 |
+
},
|
| 1427 |
+
{
|
| 1428 |
+
"epoch": 0.55,
|
| 1429 |
+
"learning_rate": 4.428373937339055e-05,
|
| 1430 |
+
"loss": 0.0,
|
| 1431 |
+
"step": 237
|
| 1432 |
+
},
|
| 1433 |
+
{
|
| 1434 |
+
"epoch": 0.55,
|
| 1435 |
+
"learning_rate": 4.391672307478739e-05,
|
| 1436 |
+
"loss": 0.0,
|
| 1437 |
+
"step": 238
|
| 1438 |
+
},
|
| 1439 |
+
{
|
| 1440 |
+
"epoch": 0.55,
|
| 1441 |
+
"learning_rate": 4.3550039173489845e-05,
|
| 1442 |
+
"loss": 0.0,
|
| 1443 |
+
"step": 239
|
| 1444 |
+
},
|
| 1445 |
+
{
|
| 1446 |
+
"epoch": 0.55,
|
| 1447 |
+
"learning_rate": 4.3183707705531496e-05,
|
| 1448 |
+
"loss": 0.0,
|
| 1449 |
+
"step": 240
|
| 1450 |
+
},
|
| 1451 |
+
{
|
| 1452 |
+
"epoch": 0.56,
|
| 1453 |
+
"learning_rate": 4.2817748687688594e-05,
|
| 1454 |
+
"loss": 0.0,
|
| 1455 |
+
"step": 241
|
| 1456 |
+
},
|
| 1457 |
+
{
|
| 1458 |
+
"epoch": 0.56,
|
| 1459 |
+
"learning_rate": 4.245218211638624e-05,
|
| 1460 |
+
"loss": 0.0,
|
| 1461 |
+
"step": 242
|
| 1462 |
+
},
|
| 1463 |
+
{
|
| 1464 |
+
"epoch": 0.56,
|
| 1465 |
+
"learning_rate": 4.2087027966605816e-05,
|
| 1466 |
+
"loss": 0.0,
|
| 1467 |
+
"step": 243
|
| 1468 |
+
},
|
| 1469 |
+
{
|
| 1470 |
+
"epoch": 0.56,
|
| 1471 |
+
"learning_rate": 4.1722306190793495e-05,
|
| 1472 |
+
"loss": 0.0,
|
| 1473 |
+
"step": 244
|
| 1474 |
+
},
|
| 1475 |
+
{
|
| 1476 |
+
"epoch": 0.56,
|
| 1477 |
+
"learning_rate": 4.135803671777011e-05,
|
| 1478 |
+
"loss": 0.0,
|
| 1479 |
+
"step": 245
|
| 1480 |
+
},
|
| 1481 |
+
{
|
| 1482 |
+
"epoch": 0.57,
|
| 1483 |
+
"learning_rate": 4.099423945164203e-05,
|
| 1484 |
+
"loss": 0.0,
|
| 1485 |
+
"step": 246
|
| 1486 |
+
},
|
| 1487 |
+
{
|
| 1488 |
+
"epoch": 0.57,
|
| 1489 |
+
"learning_rate": 4.063093427071376e-05,
|
| 1490 |
+
"loss": 0.0,
|
| 1491 |
+
"step": 247
|
| 1492 |
+
},
|
| 1493 |
+
{
|
| 1494 |
+
"epoch": 0.57,
|
| 1495 |
+
"learning_rate": 4.026814102640173e-05,
|
| 1496 |
+
"loss": 0.0,
|
| 1497 |
+
"step": 248
|
| 1498 |
+
},
|
| 1499 |
+
{
|
| 1500 |
+
"epoch": 0.57,
|
| 1501 |
+
"learning_rate": 3.99058795421495e-05,
|
| 1502 |
+
"loss": 0.0,
|
| 1503 |
+
"step": 249
|
| 1504 |
+
},
|
| 1505 |
+
{
|
| 1506 |
+
"epoch": 0.58,
|
| 1507 |
+
"learning_rate": 3.9544169612344636e-05,
|
| 1508 |
+
"loss": 0.0,
|
| 1509 |
+
"step": 250
|
| 1510 |
+
},
|
| 1511 |
+
{
|
| 1512 |
+
"epoch": 0.58,
|
| 1513 |
+
"learning_rate": 3.918303100123718e-05,
|
| 1514 |
+
"loss": 0.0,
|
| 1515 |
+
"step": 251
|
| 1516 |
+
},
|
| 1517 |
+
{
|
| 1518 |
+
"epoch": 0.58,
|
| 1519 |
+
"learning_rate": 3.8822483441859646e-05,
|
| 1520 |
+
"loss": 0.0,
|
| 1521 |
+
"step": 252
|
| 1522 |
+
},
|
| 1523 |
+
{
|
| 1524 |
+
"epoch": 0.58,
|
| 1525 |
+
"learning_rate": 3.8462546634948764e-05,
|
| 1526 |
+
"loss": 0.0,
|
| 1527 |
+
"step": 253
|
| 1528 |
+
},
|
| 1529 |
+
{
|
| 1530 |
+
"epoch": 0.59,
|
| 1531 |
+
"learning_rate": 3.8103240247869075e-05,
|
| 1532 |
+
"loss": 0.0,
|
| 1533 |
+
"step": 254
|
| 1534 |
+
},
|
| 1535 |
+
{
|
| 1536 |
+
"epoch": 0.59,
|
| 1537 |
+
"learning_rate": 3.774458391353821e-05,
|
| 1538 |
+
"loss": 0.0,
|
| 1539 |
+
"step": 255
|
| 1540 |
+
},
|
| 1541 |
+
{
|
| 1542 |
+
"epoch": 0.59,
|
| 1543 |
+
"learning_rate": 3.7386597229354225e-05,
|
| 1544 |
+
"loss": 0.0,
|
| 1545 |
+
"step": 256
|
| 1546 |
+
},
|
| 1547 |
+
{
|
| 1548 |
+
"epoch": 0.59,
|
| 1549 |
+
"learning_rate": 3.702929975612467e-05,
|
| 1550 |
+
"loss": 0.0,
|
| 1551 |
+
"step": 257
|
| 1552 |
+
},
|
| 1553 |
+
{
|
| 1554 |
+
"epoch": 0.59,
|
| 1555 |
+
"learning_rate": 3.66727110169978e-05,
|
| 1556 |
+
"loss": 0.0,
|
| 1557 |
+
"step": 258
|
| 1558 |
+
},
|
| 1559 |
+
{
|
| 1560 |
+
"epoch": 0.6,
|
| 1561 |
+
"learning_rate": 3.631685049639586e-05,
|
| 1562 |
+
"loss": 0.0,
|
| 1563 |
+
"step": 259
|
| 1564 |
+
},
|
| 1565 |
+
{
|
| 1566 |
+
"epoch": 0.6,
|
| 1567 |
+
"learning_rate": 3.5961737638950365e-05,
|
| 1568 |
+
"loss": 0.0,
|
| 1569 |
+
"step": 260
|
| 1570 |
+
},
|
| 1571 |
+
{
|
| 1572 |
+
"epoch": 0.6,
|
| 1573 |
+
"learning_rate": 3.560739184843963e-05,
|
| 1574 |
+
"loss": 0.0,
|
| 1575 |
+
"step": 261
|
| 1576 |
+
},
|
| 1577 |
+
{
|
| 1578 |
+
"epoch": 0.6,
|
| 1579 |
+
"learning_rate": 3.525383248672857e-05,
|
| 1580 |
+
"loss": 0.0,
|
| 1581 |
+
"step": 262
|
| 1582 |
+
},
|
| 1583 |
+
{
|
| 1584 |
+
"epoch": 0.61,
|
| 1585 |
+
"learning_rate": 3.4901078872710734e-05,
|
| 1586 |
+
"loss": 0.0,
|
| 1587 |
+
"step": 263
|
| 1588 |
+
},
|
| 1589 |
+
{
|
| 1590 |
+
"epoch": 0.61,
|
| 1591 |
+
"learning_rate": 3.4549150281252636e-05,
|
| 1592 |
+
"loss": 0.0,
|
| 1593 |
+
"step": 264
|
| 1594 |
+
},
|
| 1595 |
+
{
|
| 1596 |
+
"epoch": 0.61,
|
| 1597 |
+
"learning_rate": 3.419806594214063e-05,
|
| 1598 |
+
"loss": 0.0,
|
| 1599 |
+
"step": 265
|
| 1600 |
+
},
|
| 1601 |
+
{
|
| 1602 |
+
"epoch": 0.61,
|
| 1603 |
+
"learning_rate": 3.3847845039030134e-05,
|
| 1604 |
+
"loss": 0.0,
|
| 1605 |
+
"step": 266
|
| 1606 |
+
},
|
| 1607 |
+
{
|
| 1608 |
+
"epoch": 0.62,
|
| 1609 |
+
"learning_rate": 3.349850670839744e-05,
|
| 1610 |
+
"loss": 0.0,
|
| 1611 |
+
"step": 267
|
| 1612 |
+
},
|
| 1613 |
+
{
|
| 1614 |
+
"epoch": 0.62,
|
| 1615 |
+
"learning_rate": 3.315007003849403e-05,
|
| 1616 |
+
"loss": 0.0,
|
| 1617 |
+
"step": 268
|
| 1618 |
+
},
|
| 1619 |
+
{
|
| 1620 |
+
"epoch": 0.62,
|
| 1621 |
+
"learning_rate": 3.2802554068303596e-05,
|
| 1622 |
+
"loss": 0.0971,
|
| 1623 |
+
"step": 269
|
| 1624 |
+
},
|
| 1625 |
+
{
|
| 1626 |
+
"epoch": 0.62,
|
| 1627 |
+
"learning_rate": 3.245597778650176e-05,
|
| 1628 |
+
"loss": 0.0,
|
| 1629 |
+
"step": 270
|
| 1630 |
+
},
|
| 1631 |
+
{
|
| 1632 |
+
"epoch": 0.62,
|
| 1633 |
+
"learning_rate": 3.2110360130418483e-05,
|
| 1634 |
+
"loss": 0.0976,
|
| 1635 |
+
"step": 271
|
| 1636 |
+
},
|
| 1637 |
+
{
|
| 1638 |
+
"epoch": 0.63,
|
| 1639 |
+
"learning_rate": 3.176571998500325e-05,
|
| 1640 |
+
"loss": 0.0,
|
| 1641 |
+
"step": 272
|
| 1642 |
+
},
|
| 1643 |
+
{
|
| 1644 |
+
"epoch": 0.63,
|
| 1645 |
+
"learning_rate": 3.142207618179327e-05,
|
| 1646 |
+
"loss": 0.0,
|
| 1647 |
+
"step": 273
|
| 1648 |
+
},
|
| 1649 |
+
{
|
| 1650 |
+
"epoch": 0.63,
|
| 1651 |
+
"learning_rate": 3.107944749788449e-05,
|
| 1652 |
+
"loss": 0.0,
|
| 1653 |
+
"step": 274
|
| 1654 |
+
},
|
| 1655 |
+
{
|
| 1656 |
+
"epoch": 0.63,
|
| 1657 |
+
"learning_rate": 3.0737852654905497e-05,
|
| 1658 |
+
"loss": 0.0,
|
| 1659 |
+
"step": 275
|
| 1660 |
+
},
|
| 1661 |
+
{
|
| 1662 |
+
"epoch": 0.64,
|
| 1663 |
+
"learning_rate": 3.0397310317994622e-05,
|
| 1664 |
+
"loss": 0.0,
|
| 1665 |
+
"step": 276
|
| 1666 |
+
},
|
| 1667 |
+
{
|
| 1668 |
+
"epoch": 0.64,
|
| 1669 |
+
"learning_rate": 3.0057839094780106e-05,
|
| 1670 |
+
"loss": 0.0,
|
| 1671 |
+
"step": 277
|
| 1672 |
+
},
|
| 1673 |
+
{
|
| 1674 |
+
"epoch": 0.64,
|
| 1675 |
+
"learning_rate": 2.9719457534363225e-05,
|
| 1676 |
+
"loss": 0.0,
|
| 1677 |
+
"step": 278
|
| 1678 |
+
},
|
| 1679 |
+
{
|
| 1680 |
+
"epoch": 0.64,
|
| 1681 |
+
"learning_rate": 2.9382184126304834e-05,
|
| 1682 |
+
"loss": 0.0,
|
| 1683 |
+
"step": 279
|
| 1684 |
+
},
|
| 1685 |
+
{
|
| 1686 |
+
"epoch": 0.65,
|
| 1687 |
+
"learning_rate": 2.904603729961508e-05,
|
| 1688 |
+
"loss": 0.0,
|
| 1689 |
+
"step": 280
|
| 1690 |
+
},
|
| 1691 |
+
{
|
| 1692 |
+
"epoch": 0.65,
|
| 1693 |
+
"learning_rate": 2.8711035421746367e-05,
|
| 1694 |
+
"loss": 0.0,
|
| 1695 |
+
"step": 281
|
| 1696 |
+
},
|
| 1697 |
+
{
|
| 1698 |
+
"epoch": 0.65,
|
| 1699 |
+
"learning_rate": 2.837719679758981e-05,
|
| 1700 |
+
"loss": 0.0,
|
| 1701 |
+
"step": 282
|
| 1702 |
+
},
|
| 1703 |
+
{
|
| 1704 |
+
"epoch": 0.65,
|
| 1705 |
+
"learning_rate": 2.804453966847497e-05,
|
| 1706 |
+
"loss": 0.0,
|
| 1707 |
+
"step": 283
|
| 1708 |
+
},
|
| 1709 |
+
{
|
| 1710 |
+
"epoch": 0.65,
|
| 1711 |
+
"learning_rate": 2.771308221117309e-05,
|
| 1712 |
+
"loss": 0.0,
|
| 1713 |
+
"step": 284
|
| 1714 |
+
},
|
| 1715 |
+
{
|
| 1716 |
+
"epoch": 0.66,
|
| 1717 |
+
"learning_rate": 2.7382842536904075e-05,
|
| 1718 |
+
"loss": 0.0,
|
| 1719 |
+
"step": 285
|
| 1720 |
+
},
|
| 1721 |
+
{
|
| 1722 |
+
"epoch": 0.66,
|
| 1723 |
+
"learning_rate": 2.705383869034664e-05,
|
| 1724 |
+
"loss": 0.0,
|
| 1725 |
+
"step": 286
|
| 1726 |
+
},
|
| 1727 |
+
{
|
| 1728 |
+
"epoch": 0.66,
|
| 1729 |
+
"learning_rate": 2.6726088648652503e-05,
|
| 1730 |
+
"loss": 0.0,
|
| 1731 |
+
"step": 287
|
| 1732 |
+
},
|
| 1733 |
+
{
|
| 1734 |
+
"epoch": 0.66,
|
| 1735 |
+
"learning_rate": 2.639961032046404e-05,
|
| 1736 |
+
"loss": 0.0172,
|
| 1737 |
+
"step": 288
|
| 1738 |
+
},
|
| 1739 |
+
{
|
| 1740 |
+
"epoch": 0.67,
|
| 1741 |
+
"learning_rate": 2.607442154493568e-05,
|
| 1742 |
+
"loss": 0.0,
|
| 1743 |
+
"step": 289
|
| 1744 |
+
},
|
| 1745 |
+
{
|
| 1746 |
+
"epoch": 0.67,
|
| 1747 |
+
"learning_rate": 2.5750540090759224e-05,
|
| 1748 |
+
"loss": 0.0252,
|
| 1749 |
+
"step": 290
|
| 1750 |
+
},
|
| 1751 |
+
{
|
| 1752 |
+
"epoch": 0.67,
|
| 1753 |
+
"learning_rate": 2.5427983655192888e-05,
|
| 1754 |
+
"loss": 0.0015,
|
| 1755 |
+
"step": 291
|
| 1756 |
+
},
|
| 1757 |
+
{
|
| 1758 |
+
"epoch": 0.67,
|
| 1759 |
+
"learning_rate": 2.5106769863094386e-05,
|
| 1760 |
+
"loss": 0.0,
|
| 1761 |
+
"step": 292
|
| 1762 |
+
},
|
| 1763 |
+
{
|
| 1764 |
+
"epoch": 0.68,
|
| 1765 |
+
"learning_rate": 2.4786916265957834e-05,
|
| 1766 |
+
"loss": 0.0,
|
| 1767 |
+
"step": 293
|
| 1768 |
+
},
|
| 1769 |
+
{
|
| 1770 |
+
"epoch": 0.68,
|
| 1771 |
+
"learning_rate": 2.446844034095466e-05,
|
| 1772 |
+
"loss": 0.0,
|
| 1773 |
+
"step": 294
|
| 1774 |
+
},
|
| 1775 |
+
{
|
| 1776 |
+
"epoch": 0.68,
|
| 1777 |
+
"learning_rate": 2.4151359489978733e-05,
|
| 1778 |
+
"loss": 0.0,
|
| 1779 |
+
"step": 295
|
| 1780 |
+
},
|
| 1781 |
+
{
|
| 1782 |
+
"epoch": 0.68,
|
| 1783 |
+
"learning_rate": 2.3835691038695462e-05,
|
| 1784 |
+
"loss": 0.0001,
|
| 1785 |
+
"step": 296
|
| 1786 |
+
},
|
| 1787 |
+
{
|
| 1788 |
+
"epoch": 0.68,
|
| 1789 |
+
"learning_rate": 2.3521452235595066e-05,
|
| 1790 |
+
"loss": 0.0001,
|
| 1791 |
+
"step": 297
|
| 1792 |
+
},
|
| 1793 |
+
{
|
| 1794 |
+
"epoch": 0.69,
|
| 1795 |
+
"learning_rate": 2.3208660251050158e-05,
|
| 1796 |
+
"loss": 0.0001,
|
| 1797 |
+
"step": 298
|
| 1798 |
+
},
|
| 1799 |
+
{
|
| 1800 |
+
"epoch": 0.69,
|
| 1801 |
+
"learning_rate": 2.2897332176377528e-05,
|
| 1802 |
+
"loss": 0.0,
|
| 1803 |
+
"step": 299
|
| 1804 |
+
},
|
| 1805 |
+
{
|
| 1806 |
+
"epoch": 0.69,
|
| 1807 |
+
"learning_rate": 2.2587485022904187e-05,
|
| 1808 |
+
"loss": 0.0,
|
| 1809 |
+
"step": 300
|
| 1810 |
+
},
|
| 1811 |
+
{
|
| 1812 |
+
"epoch": 0.69,
|
| 1813 |
+
"learning_rate": 2.2279135721037908e-05,
|
| 1814 |
+
"loss": 0.0,
|
| 1815 |
+
"step": 301
|
| 1816 |
+
},
|
| 1817 |
+
{
|
| 1818 |
+
"epoch": 0.7,
|
| 1819 |
+
"learning_rate": 2.197230111934211e-05,
|
| 1820 |
+
"loss": 0.0,
|
| 1821 |
+
"step": 302
|
| 1822 |
+
},
|
| 1823 |
+
{
|
| 1824 |
+
"epoch": 0.7,
|
| 1825 |
+
"learning_rate": 2.166699798361527e-05,
|
| 1826 |
+
"loss": 0.0,
|
| 1827 |
+
"step": 303
|
| 1828 |
+
},
|
| 1829 |
+
{
|
| 1830 |
+
"epoch": 0.7,
|
| 1831 |
+
"learning_rate": 2.136324299597474e-05,
|
| 1832 |
+
"loss": 0.0,
|
| 1833 |
+
"step": 304
|
| 1834 |
+
},
|
| 1835 |
+
{
|
| 1836 |
+
"epoch": 0.7,
|
| 1837 |
+
"learning_rate": 2.106105275394533e-05,
|
| 1838 |
+
"loss": 0.0,
|
| 1839 |
+
"step": 305
|
| 1840 |
+
},
|
| 1841 |
+
{
|
| 1842 |
+
"epoch": 0.71,
|
| 1843 |
+
"learning_rate": 2.0760443769552258e-05,
|
| 1844 |
+
"loss": 0.0,
|
| 1845 |
+
"step": 306
|
| 1846 |
+
},
|
| 1847 |
+
{
|
| 1848 |
+
"epoch": 0.71,
|
| 1849 |
+
"learning_rate": 2.046143246841906e-05,
|
| 1850 |
+
"loss": 0.0,
|
| 1851 |
+
"step": 307
|
| 1852 |
+
},
|
| 1853 |
+
{
|
| 1854 |
+
"epoch": 0.71,
|
| 1855 |
+
"learning_rate": 2.0164035188869968e-05,
|
| 1856 |
+
"loss": 0.0,
|
| 1857 |
+
"step": 308
|
| 1858 |
+
},
|
| 1859 |
+
{
|
| 1860 |
+
"epoch": 0.71,
|
| 1861 |
+
"learning_rate": 1.9868268181037185e-05,
|
| 1862 |
+
"loss": 0.0,
|
| 1863 |
+
"step": 309
|
| 1864 |
+
},
|
| 1865 |
+
{
|
| 1866 |
+
"epoch": 0.71,
|
| 1867 |
+
"learning_rate": 1.957414760597303e-05,
|
| 1868 |
+
"loss": 0.0,
|
| 1869 |
+
"step": 310
|
| 1870 |
+
},
|
| 1871 |
+
{
|
| 1872 |
+
"epoch": 0.72,
|
| 1873 |
+
"learning_rate": 1.9281689534766813e-05,
|
| 1874 |
+
"loss": 0.0,
|
| 1875 |
+
"step": 311
|
| 1876 |
+
},
|
| 1877 |
+
{
|
| 1878 |
+
"epoch": 0.72,
|
| 1879 |
+
"learning_rate": 1.8990909947666668e-05,
|
| 1880 |
+
"loss": 0.0,
|
| 1881 |
+
"step": 312
|
| 1882 |
+
},
|
| 1883 |
+
{
|
| 1884 |
+
"epoch": 0.72,
|
| 1885 |
+
"learning_rate": 1.870182473320642e-05,
|
| 1886 |
+
"loss": 0.0,
|
| 1887 |
+
"step": 313
|
| 1888 |
+
},
|
| 1889 |
+
{
|
| 1890 |
+
"epoch": 0.72,
|
| 1891 |
+
"learning_rate": 1.8414449687337464e-05,
|
| 1892 |
+
"loss": 0.0,
|
| 1893 |
+
"step": 314
|
| 1894 |
+
},
|
| 1895 |
+
{
|
| 1896 |
+
"epoch": 0.73,
|
| 1897 |
+
"learning_rate": 1.8128800512565513e-05,
|
| 1898 |
+
"loss": 0.0,
|
| 1899 |
+
"step": 315
|
| 1900 |
+
},
|
| 1901 |
+
{
|
| 1902 |
+
"epoch": 0.73,
|
| 1903 |
+
"learning_rate": 1.7844892817092756e-05,
|
| 1904 |
+
"loss": 0.0,
|
| 1905 |
+
"step": 316
|
| 1906 |
+
},
|
| 1907 |
+
{
|
| 1908 |
+
"epoch": 0.73,
|
| 1909 |
+
"learning_rate": 1.7562742113964914e-05,
|
| 1910 |
+
"loss": 0.0,
|
| 1911 |
+
"step": 317
|
| 1912 |
+
},
|
| 1913 |
+
{
|
| 1914 |
+
"epoch": 0.73,
|
| 1915 |
+
"learning_rate": 1.728236382022359e-05,
|
| 1916 |
+
"loss": 0.0,
|
| 1917 |
+
"step": 318
|
| 1918 |
+
},
|
| 1919 |
+
{
|
| 1920 |
+
"epoch": 0.74,
|
| 1921 |
+
"learning_rate": 1.700377325606388e-05,
|
| 1922 |
+
"loss": 0.0,
|
| 1923 |
+
"step": 319
|
| 1924 |
+
},
|
| 1925 |
+
{
|
| 1926 |
+
"epoch": 0.74,
|
| 1927 |
+
"learning_rate": 1.6726985643997273e-05,
|
| 1928 |
+
"loss": 0.0,
|
| 1929 |
+
"step": 320
|
| 1930 |
+
},
|
| 1931 |
+
{
|
| 1932 |
+
"epoch": 0.74,
|
| 1933 |
+
"learning_rate": 1.645201610801988e-05,
|
| 1934 |
+
"loss": 0.0,
|
| 1935 |
+
"step": 321
|
| 1936 |
+
},
|
| 1937 |
+
{
|
| 1938 |
+
"epoch": 0.74,
|
| 1939 |
+
"learning_rate": 1.617887967278602e-05,
|
| 1940 |
+
"loss": 0.0,
|
| 1941 |
+
"step": 322
|
| 1942 |
+
},
|
| 1943 |
+
{
|
| 1944 |
+
"epoch": 0.74,
|
| 1945 |
+
"learning_rate": 1.5907591262787247e-05,
|
| 1946 |
+
"loss": 0.0,
|
| 1947 |
+
"step": 323
|
| 1948 |
+
},
|
| 1949 |
+
{
|
| 1950 |
+
"epoch": 0.75,
|
| 1951 |
+
"learning_rate": 1.5638165701536868e-05,
|
| 1952 |
+
"loss": 0.0,
|
| 1953 |
+
"step": 324
|
| 1954 |
+
},
|
| 1955 |
+
{
|
| 1956 |
+
"epoch": 0.75,
|
| 1957 |
+
"learning_rate": 1.5370617710760004e-05,
|
| 1958 |
+
"loss": 0.0221,
|
| 1959 |
+
"step": 325
|
| 1960 |
+
},
|
| 1961 |
+
{
|
| 1962 |
+
"epoch": 0.75,
|
| 1963 |
+
"learning_rate": 1.5104961909589116e-05,
|
| 1964 |
+
"loss": 0.0,
|
| 1965 |
+
"step": 326
|
| 1966 |
+
},
|
| 1967 |
+
{
|
| 1968 |
+
"epoch": 0.75,
|
| 1969 |
+
"learning_rate": 1.4841212813765255e-05,
|
| 1970 |
+
"loss": 0.0,
|
| 1971 |
+
"step": 327
|
| 1972 |
+
},
|
| 1973 |
+
{
|
| 1974 |
+
"epoch": 0.76,
|
| 1975 |
+
"learning_rate": 1.457938483484489e-05,
|
| 1976 |
+
"loss": 0.0,
|
| 1977 |
+
"step": 328
|
| 1978 |
+
},
|
| 1979 |
+
{
|
| 1980 |
+
"epoch": 0.76,
|
| 1981 |
+
"learning_rate": 1.4319492279412388e-05,
|
| 1982 |
+
"loss": 0.0,
|
| 1983 |
+
"step": 329
|
| 1984 |
+
},
|
| 1985 |
+
{
|
| 1986 |
+
"epoch": 0.76,
|
| 1987 |
+
"learning_rate": 1.4061549348298376e-05,
|
| 1988 |
+
"loss": 0.0,
|
| 1989 |
+
"step": 330
|
| 1990 |
+
},
|
| 1991 |
+
{
|
| 1992 |
+
"epoch": 0.76,
|
| 1993 |
+
"learning_rate": 1.3805570135803708e-05,
|
| 1994 |
+
"loss": 0.0,
|
| 1995 |
+
"step": 331
|
| 1996 |
+
},
|
| 1997 |
+
{
|
| 1998 |
+
"epoch": 0.76,
|
| 1999 |
+
"learning_rate": 1.3551568628929434e-05,
|
| 2000 |
+
"loss": 0.0,
|
| 2001 |
+
"step": 332
|
| 2002 |
+
},
|
| 2003 |
+
{
|
| 2004 |
+
"epoch": 0.77,
|
| 2005 |
+
"learning_rate": 1.3299558706612409e-05,
|
| 2006 |
+
"loss": 0.0,
|
| 2007 |
+
"step": 333
|
| 2008 |
+
},
|
| 2009 |
+
{
|
| 2010 |
+
"epoch": 0.77,
|
| 2011 |
+
"learning_rate": 1.3049554138967051e-05,
|
| 2012 |
+
"loss": 0.0,
|
| 2013 |
+
"step": 334
|
| 2014 |
+
},
|
| 2015 |
+
{
|
| 2016 |
+
"epoch": 0.77,
|
| 2017 |
+
"learning_rate": 1.2801568586532837e-05,
|
| 2018 |
+
"loss": 0.0,
|
| 2019 |
+
"step": 335
|
| 2020 |
+
},
|
| 2021 |
+
{
|
| 2022 |
+
"epoch": 0.77,
|
| 2023 |
+
"learning_rate": 1.2555615599527948e-05,
|
| 2024 |
+
"loss": 0.0,
|
| 2025 |
+
"step": 336
|
| 2026 |
+
},
|
| 2027 |
+
{
|
| 2028 |
+
"epoch": 0.78,
|
| 2029 |
+
"learning_rate": 1.231170861710879e-05,
|
| 2030 |
+
"loss": 0.0,
|
| 2031 |
+
"step": 337
|
| 2032 |
+
},
|
| 2033 |
+
{
|
| 2034 |
+
"epoch": 0.78,
|
| 2035 |
+
"learning_rate": 1.2069860966635732e-05,
|
| 2036 |
+
"loss": 0.0,
|
| 2037 |
+
"step": 338
|
| 2038 |
+
},
|
| 2039 |
+
{
|
| 2040 |
+
"epoch": 0.78,
|
| 2041 |
+
"learning_rate": 1.183008586294485e-05,
|
| 2042 |
+
"loss": 0.0,
|
| 2043 |
+
"step": 339
|
| 2044 |
+
},
|
| 2045 |
+
{
|
| 2046 |
+
"epoch": 0.78,
|
| 2047 |
+
"learning_rate": 1.1592396407625883e-05,
|
| 2048 |
+
"loss": 0.0,
|
| 2049 |
+
"step": 340
|
| 2050 |
+
},
|
| 2051 |
+
{
|
| 2052 |
+
"epoch": 0.79,
|
| 2053 |
+
"learning_rate": 1.1356805588306291e-05,
|
| 2054 |
+
"loss": 0.0209,
|
| 2055 |
+
"step": 341
|
| 2056 |
+
},
|
| 2057 |
+
{
|
| 2058 |
+
"epoch": 0.79,
|
| 2059 |
+
"learning_rate": 1.1123326277941626e-05,
|
| 2060 |
+
"loss": 0.0,
|
| 2061 |
+
"step": 342
|
| 2062 |
+
},
|
| 2063 |
+
{
|
| 2064 |
+
"epoch": 0.79,
|
| 2065 |
+
"learning_rate": 1.0891971234112191e-05,
|
| 2066 |
+
"loss": 0.0,
|
| 2067 |
+
"step": 343
|
| 2068 |
+
},
|
| 2069 |
+
{
|
| 2070 |
+
"epoch": 0.79,
|
| 2071 |
+
"learning_rate": 1.066275309832584e-05,
|
| 2072 |
+
"loss": 0.0,
|
| 2073 |
+
"step": 344
|
| 2074 |
+
},
|
| 2075 |
+
{
|
| 2076 |
+
"epoch": 0.79,
|
| 2077 |
+
"learning_rate": 1.043568439532735e-05,
|
| 2078 |
+
"loss": 0.0124,
|
| 2079 |
+
"step": 345
|
| 2080 |
+
},
|
| 2081 |
+
{
|
| 2082 |
+
"epoch": 0.8,
|
| 2083 |
+
"learning_rate": 1.0210777532413939e-05,
|
| 2084 |
+
"loss": 0.0,
|
| 2085 |
+
"step": 346
|
| 2086 |
+
},
|
| 2087 |
+
{
|
| 2088 |
+
"epoch": 0.8,
|
| 2089 |
+
"learning_rate": 9.988044798757452e-06,
|
| 2090 |
+
"loss": 0.0,
|
| 2091 |
+
"step": 347
|
| 2092 |
+
},
|
| 2093 |
+
{
|
| 2094 |
+
"epoch": 0.8,
|
| 2095 |
+
"learning_rate": 9.767498364732713e-06,
|
| 2096 |
+
"loss": 0.0,
|
| 2097 |
+
"step": 348
|
| 2098 |
+
},
|
| 2099 |
+
{
|
| 2100 |
+
"epoch": 0.8,
|
| 2101 |
+
"learning_rate": 9.549150281252633e-06,
|
| 2102 |
+
"loss": 0.0,
|
| 2103 |
+
"step": 349
|
| 2104 |
+
},
|
| 2105 |
+
{
|
| 2106 |
+
"epoch": 0.81,
|
| 2107 |
+
"learning_rate": 9.333012479109704e-06,
|
| 2108 |
+
"loss": 0.0,
|
| 2109 |
+
"step": 350
|
| 2110 |
+
},
|
| 2111 |
+
{
|
| 2112 |
+
"epoch": 0.81,
|
| 2113 |
+
"learning_rate": 9.119096768324097e-06,
|
| 2114 |
+
"loss": 0.0,
|
| 2115 |
+
"step": 351
|
| 2116 |
+
},
|
| 2117 |
+
{
|
| 2118 |
+
"epoch": 0.81,
|
| 2119 |
+
"learning_rate": 8.907414837498296e-06,
|
| 2120 |
+
"loss": 0.0,
|
| 2121 |
+
"step": 352
|
| 2122 |
+
},
|
| 2123 |
+
{
|
| 2124 |
+
"epoch": 0.81,
|
| 2125 |
+
"learning_rate": 8.697978253178468e-06,
|
| 2126 |
+
"loss": 0.0,
|
| 2127 |
+
"step": 353
|
| 2128 |
+
},
|
| 2129 |
+
{
|
| 2130 |
+
"epoch": 0.82,
|
| 2131 |
+
"learning_rate": 8.490798459222476e-06,
|
| 2132 |
+
"loss": 0.0,
|
| 2133 |
+
"step": 354
|
| 2134 |
+
},
|
| 2135 |
+
{
|
| 2136 |
+
"epoch": 0.82,
|
| 2137 |
+
"learning_rate": 8.2858867761745e-06,
|
| 2138 |
+
"loss": 0.0,
|
| 2139 |
+
"step": 355
|
| 2140 |
+
},
|
| 2141 |
+
{
|
| 2142 |
+
"epoch": 0.82,
|
| 2143 |
+
"learning_rate": 8.083254400646506e-06,
|
| 2144 |
+
"loss": 0.0,
|
| 2145 |
+
"step": 356
|
| 2146 |
+
},
|
| 2147 |
+
{
|
| 2148 |
+
"epoch": 0.82,
|
| 2149 |
+
"learning_rate": 7.882912404706528e-06,
|
| 2150 |
+
"loss": 0.0,
|
| 2151 |
+
"step": 357
|
| 2152 |
+
},
|
| 2153 |
+
{
|
| 2154 |
+
"epoch": 0.82,
|
| 2155 |
+
"learning_rate": 7.684871735273513e-06,
|
| 2156 |
+
"loss": 0.0,
|
| 2157 |
+
"step": 358
|
| 2158 |
+
},
|
| 2159 |
+
{
|
| 2160 |
+
"epoch": 0.83,
|
| 2161 |
+
"learning_rate": 7.489143213519301e-06,
|
| 2162 |
+
"loss": 0.0,
|
| 2163 |
+
"step": 359
|
| 2164 |
+
},
|
| 2165 |
+
{
|
| 2166 |
+
"epoch": 0.83,
|
| 2167 |
+
"learning_rate": 7.295737534277286e-06,
|
| 2168 |
+
"loss": 0.0,
|
| 2169 |
+
"step": 360
|
| 2170 |
+
},
|
| 2171 |
+
{
|
| 2172 |
+
"epoch": 0.83,
|
| 2173 |
+
"learning_rate": 7.104665265458083e-06,
|
| 2174 |
+
"loss": 0.0,
|
| 2175 |
+
"step": 361
|
| 2176 |
+
},
|
| 2177 |
+
{
|
| 2178 |
+
"epoch": 0.83,
|
| 2179 |
+
"learning_rate": 6.9159368474720155e-06,
|
| 2180 |
+
"loss": 0.0,
|
| 2181 |
+
"step": 362
|
| 2182 |
+
},
|
| 2183 |
+
{
|
| 2184 |
+
"epoch": 0.84,
|
| 2185 |
+
"learning_rate": 6.729562592658717e-06,
|
| 2186 |
+
"loss": 0.0,
|
| 2187 |
+
"step": 363
|
| 2188 |
+
},
|
| 2189 |
+
{
|
| 2190 |
+
"epoch": 0.84,
|
| 2191 |
+
"learning_rate": 6.5455526847235825e-06,
|
| 2192 |
+
"loss": 0.0,
|
| 2193 |
+
"step": 364
|
| 2194 |
+
},
|
| 2195 |
+
{
|
| 2196 |
+
"epoch": 0.84,
|
| 2197 |
+
"learning_rate": 6.363917178181389e-06,
|
| 2198 |
+
"loss": 0.0,
|
| 2199 |
+
"step": 365
|
| 2200 |
+
},
|
| 2201 |
+
{
|
| 2202 |
+
"epoch": 0.84,
|
| 2203 |
+
"learning_rate": 6.184665997806832e-06,
|
| 2204 |
+
"loss": 0.0,
|
| 2205 |
+
"step": 366
|
| 2206 |
+
},
|
| 2207 |
+
{
|
| 2208 |
+
"epoch": 0.85,
|
| 2209 |
+
"learning_rate": 6.007808938092263e-06,
|
| 2210 |
+
"loss": 0.0,
|
| 2211 |
+
"step": 367
|
| 2212 |
+
},
|
| 2213 |
+
{
|
| 2214 |
+
"epoch": 0.85,
|
| 2215 |
+
"learning_rate": 5.833355662712525e-06,
|
| 2216 |
+
"loss": 0.0,
|
| 2217 |
+
"step": 368
|
| 2218 |
+
},
|
| 2219 |
+
{
|
| 2220 |
+
"epoch": 0.85,
|
| 2221 |
+
"learning_rate": 5.6613157039969055e-06,
|
| 2222 |
+
"loss": 0.0,
|
| 2223 |
+
"step": 369
|
| 2224 |
+
},
|
| 2225 |
+
{
|
| 2226 |
+
"epoch": 0.85,
|
| 2227 |
+
"learning_rate": 5.49169846240824e-06,
|
| 2228 |
+
"loss": 0.0,
|
| 2229 |
+
"step": 370
|
| 2230 |
+
},
|
| 2231 |
+
{
|
| 2232 |
+
"epoch": 0.85,
|
| 2233 |
+
"learning_rate": 5.3245132060292915e-06,
|
| 2234 |
+
"loss": 0.0,
|
| 2235 |
+
"step": 371
|
| 2236 |
+
},
|
| 2237 |
+
{
|
| 2238 |
+
"epoch": 0.86,
|
| 2239 |
+
"learning_rate": 5.159769070056364e-06,
|
| 2240 |
+
"loss": 0.0,
|
| 2241 |
+
"step": 372
|
| 2242 |
+
},
|
| 2243 |
+
{
|
| 2244 |
+
"epoch": 0.86,
|
| 2245 |
+
"learning_rate": 4.997475056300077e-06,
|
| 2246 |
+
"loss": 0.0,
|
| 2247 |
+
"step": 373
|
| 2248 |
+
},
|
| 2249 |
+
{
|
| 2250 |
+
"epoch": 0.86,
|
| 2251 |
+
"learning_rate": 4.837640032693558e-06,
|
| 2252 |
+
"loss": 0.0,
|
| 2253 |
+
"step": 374
|
| 2254 |
+
},
|
| 2255 |
+
{
|
| 2256 |
+
"epoch": 0.86,
|
| 2257 |
+
"learning_rate": 4.680272732807834e-06,
|
| 2258 |
+
"loss": 0.0,
|
| 2259 |
+
"step": 375
|
| 2260 |
+
},
|
| 2261 |
+
{
|
| 2262 |
+
"epoch": 0.87,
|
| 2263 |
+
"learning_rate": 4.52538175537468e-06,
|
| 2264 |
+
"loss": 0.0,
|
| 2265 |
+
"step": 376
|
| 2266 |
+
},
|
| 2267 |
+
{
|
| 2268 |
+
"epoch": 0.87,
|
| 2269 |
+
"learning_rate": 4.372975563816706e-06,
|
| 2270 |
+
"loss": 0.0,
|
| 2271 |
+
"step": 377
|
| 2272 |
+
},
|
| 2273 |
+
{
|
| 2274 |
+
"epoch": 0.87,
|
| 2275 |
+
"learning_rate": 4.223062485784945e-06,
|
| 2276 |
+
"loss": 0.0,
|
| 2277 |
+
"step": 378
|
| 2278 |
+
},
|
| 2279 |
+
{
|
| 2280 |
+
"epoch": 0.87,
|
| 2281 |
+
"learning_rate": 4.075650712703849e-06,
|
| 2282 |
+
"loss": 0.0,
|
| 2283 |
+
"step": 379
|
| 2284 |
+
},
|
| 2285 |
+
{
|
| 2286 |
+
"epoch": 0.88,
|
| 2287 |
+
"learning_rate": 3.930748299323639e-06,
|
| 2288 |
+
"loss": 0.0,
|
| 2289 |
+
"step": 380
|
| 2290 |
+
},
|
| 2291 |
+
{
|
| 2292 |
+
"epoch": 0.88,
|
| 2293 |
+
"learning_rate": 3.788363163280212e-06,
|
| 2294 |
+
"loss": 0.0,
|
| 2295 |
+
"step": 381
|
| 2296 |
+
},
|
| 2297 |
+
{
|
| 2298 |
+
"epoch": 0.88,
|
| 2299 |
+
"learning_rate": 3.6485030846625077e-06,
|
| 2300 |
+
"loss": 0.0,
|
| 2301 |
+
"step": 382
|
| 2302 |
+
},
|
| 2303 |
+
{
|
| 2304 |
+
"epoch": 0.88,
|
| 2305 |
+
"learning_rate": 3.511175705587433e-06,
|
| 2306 |
+
"loss": 0.0,
|
| 2307 |
+
"step": 383
|
| 2308 |
+
},
|
| 2309 |
+
{
|
| 2310 |
+
"epoch": 0.88,
|
| 2311 |
+
"learning_rate": 3.376388529782215e-06,
|
| 2312 |
+
"loss": 0.0,
|
| 2313 |
+
"step": 384
|
| 2314 |
+
},
|
| 2315 |
+
{
|
| 2316 |
+
"epoch": 0.89,
|
| 2317 |
+
"learning_rate": 3.244148922174445e-06,
|
| 2318 |
+
"loss": 0.0,
|
| 2319 |
+
"step": 385
|
| 2320 |
+
},
|
| 2321 |
+
{
|
| 2322 |
+
"epoch": 0.89,
|
| 2323 |
+
"learning_rate": 3.1144641084896285e-06,
|
| 2324 |
+
"loss": 0.0,
|
| 2325 |
+
"step": 386
|
| 2326 |
+
},
|
| 2327 |
+
{
|
| 2328 |
+
"epoch": 0.89,
|
| 2329 |
+
"learning_rate": 2.987341174856395e-06,
|
| 2330 |
+
"loss": 0.0,
|
| 2331 |
+
"step": 387
|
| 2332 |
+
},
|
| 2333 |
+
{
|
| 2334 |
+
"epoch": 0.89,
|
| 2335 |
+
"learning_rate": 2.862787067419237e-06,
|
| 2336 |
+
"loss": 0.0,
|
| 2337 |
+
"step": 388
|
| 2338 |
+
},
|
| 2339 |
+
{
|
| 2340 |
+
"epoch": 0.9,
|
| 2341 |
+
"learning_rate": 2.7408085919590264e-06,
|
| 2342 |
+
"loss": 0.0,
|
| 2343 |
+
"step": 389
|
| 2344 |
+
},
|
| 2345 |
+
{
|
| 2346 |
+
"epoch": 0.9,
|
| 2347 |
+
"learning_rate": 2.6214124135211184e-06,
|
| 2348 |
+
"loss": 0.0,
|
| 2349 |
+
"step": 390
|
| 2350 |
+
},
|
| 2351 |
+
{
|
| 2352 |
+
"epoch": 0.9,
|
| 2353 |
+
"learning_rate": 2.504605056051146e-06,
|
| 2354 |
+
"loss": 0.0,
|
| 2355 |
+
"step": 391
|
| 2356 |
+
},
|
| 2357 |
+
{
|
| 2358 |
+
"epoch": 0.9,
|
| 2359 |
+
"learning_rate": 2.3903929020385962e-06,
|
| 2360 |
+
"loss": 0.0,
|
| 2361 |
+
"step": 392
|
| 2362 |
+
},
|
| 2363 |
+
{
|
| 2364 |
+
"epoch": 0.91,
|
| 2365 |
+
"learning_rate": 2.2787821921679885e-06,
|
| 2366 |
+
"loss": 0.0,
|
| 2367 |
+
"step": 393
|
| 2368 |
+
},
|
| 2369 |
+
{
|
| 2370 |
+
"epoch": 0.91,
|
| 2371 |
+
"learning_rate": 2.1697790249779636e-06,
|
| 2372 |
+
"loss": 0.0,
|
| 2373 |
+
"step": 394
|
| 2374 |
+
},
|
| 2375 |
+
{
|
| 2376 |
+
"epoch": 0.91,
|
| 2377 |
+
"learning_rate": 2.0633893565279615e-06,
|
| 2378 |
+
"loss": 0.0,
|
| 2379 |
+
"step": 395
|
| 2380 |
+
},
|
| 2381 |
+
{
|
| 2382 |
+
"epoch": 0.91,
|
| 2383 |
+
"learning_rate": 1.9596190000728433e-06,
|
| 2384 |
+
"loss": 0.0,
|
| 2385 |
+
"step": 396
|
| 2386 |
+
},
|
| 2387 |
+
{
|
| 2388 |
+
"epoch": 0.91,
|
| 2389 |
+
"learning_rate": 1.8584736257452272e-06,
|
| 2390 |
+
"loss": 0.0,
|
| 2391 |
+
"step": 397
|
| 2392 |
+
},
|
| 2393 |
+
{
|
| 2394 |
+
"epoch": 0.92,
|
| 2395 |
+
"learning_rate": 1.7599587602456559e-06,
|
| 2396 |
+
"loss": 0.0,
|
| 2397 |
+
"step": 398
|
| 2398 |
+
},
|
| 2399 |
+
{
|
| 2400 |
+
"epoch": 0.92,
|
| 2401 |
+
"learning_rate": 1.6640797865406288e-06,
|
| 2402 |
+
"loss": 0.0,
|
| 2403 |
+
"step": 399
|
| 2404 |
+
},
|
| 2405 |
+
{
|
| 2406 |
+
"epoch": 0.92,
|
| 2407 |
+
"learning_rate": 1.5708419435684462e-06,
|
| 2408 |
+
"loss": 0.0,
|
| 2409 |
+
"step": 400
|
| 2410 |
+
},
|
| 2411 |
+
{
|
| 2412 |
+
"epoch": 0.92,
|
| 2413 |
+
"learning_rate": 1.4802503259529753e-06,
|
| 2414 |
+
"loss": 0.0,
|
| 2415 |
+
"step": 401
|
| 2416 |
+
},
|
| 2417 |
+
{
|
| 2418 |
+
"epoch": 0.93,
|
| 2419 |
+
"learning_rate": 1.3923098837252457e-06,
|
| 2420 |
+
"loss": 0.0,
|
| 2421 |
+
"step": 402
|
| 2422 |
+
},
|
| 2423 |
+
{
|
| 2424 |
+
"epoch": 0.93,
|
| 2425 |
+
"learning_rate": 1.307025422053021e-06,
|
| 2426 |
+
"loss": 0.0,
|
| 2427 |
+
"step": 403
|
| 2428 |
+
},
|
| 2429 |
+
{
|
| 2430 |
+
"epoch": 0.93,
|
| 2431 |
+
"learning_rate": 1.2244016009781701e-06,
|
| 2432 |
+
"loss": 0.0,
|
| 2433 |
+
"step": 404
|
| 2434 |
+
},
|
| 2435 |
+
{
|
| 2436 |
+
"epoch": 0.93,
|
| 2437 |
+
"learning_rate": 1.1444429351620933e-06,
|
| 2438 |
+
"loss": 0.0,
|
| 2439 |
+
"step": 405
|
| 2440 |
+
},
|
| 2441 |
+
{
|
| 2442 |
+
"epoch": 0.94,
|
| 2443 |
+
"learning_rate": 1.0671537936390198e-06,
|
| 2444 |
+
"loss": 0.0,
|
| 2445 |
+
"step": 406
|
| 2446 |
+
},
|
| 2447 |
+
{
|
| 2448 |
+
"epoch": 0.94,
|
| 2449 |
+
"learning_rate": 9.92538399577264e-07,
|
| 2450 |
+
"loss": 0.0,
|
| 2451 |
+
"step": 407
|
| 2452 |
+
},
|
| 2453 |
+
{
|
| 2454 |
+
"epoch": 0.94,
|
| 2455 |
+
"learning_rate": 9.206008300484848e-07,
|
| 2456 |
+
"loss": 0.0,
|
| 2457 |
+
"step": 408
|
| 2458 |
+
},
|
| 2459 |
+
{
|
| 2460 |
+
"epoch": 0.94,
|
| 2461 |
+
"learning_rate": 8.513450158049108e-07,
|
| 2462 |
+
"loss": 0.0,
|
| 2463 |
+
"step": 409
|
| 2464 |
+
},
|
| 2465 |
+
{
|
| 2466 |
+
"epoch": 0.94,
|
| 2467 |
+
"learning_rate": 7.847747410645368e-07,
|
| 2468 |
+
"loss": 0.0,
|
| 2469 |
+
"step": 410
|
| 2470 |
+
},
|
| 2471 |
+
{
|
| 2472 |
+
"epoch": 0.95,
|
| 2473 |
+
"learning_rate": 7.208936433043767e-07,
|
| 2474 |
+
"loss": 0.0,
|
| 2475 |
+
"step": 411
|
| 2476 |
+
},
|
| 2477 |
+
{
|
| 2478 |
+
"epoch": 0.95,
|
| 2479 |
+
"learning_rate": 6.597052130616899e-07,
|
| 2480 |
+
"loss": 0.0,
|
| 2481 |
+
"step": 412
|
| 2482 |
+
},
|
| 2483 |
+
{
|
| 2484 |
+
"epoch": 0.95,
|
| 2485 |
+
"learning_rate": 6.012127937432554e-07,
|
| 2486 |
+
"loss": 0.0,
|
| 2487 |
+
"step": 413
|
| 2488 |
+
},
|
| 2489 |
+
{
|
| 2490 |
+
"epoch": 0.95,
|
| 2491 |
+
"learning_rate": 5.454195814427021e-07,
|
| 2492 |
+
"loss": 0.0,
|
| 2493 |
+
"step": 414
|
| 2494 |
+
},
|
| 2495 |
+
{
|
| 2496 |
+
"epoch": 0.96,
|
| 2497 |
+
"learning_rate": 4.923286247658476e-07,
|
| 2498 |
+
"loss": 0.0,
|
| 2499 |
+
"step": 415
|
| 2500 |
+
},
|
| 2501 |
+
{
|
| 2502 |
+
"epoch": 0.96,
|
| 2503 |
+
"learning_rate": 4.419428246641377e-07,
|
| 2504 |
+
"loss": 0.0,
|
| 2505 |
+
"step": 416
|
| 2506 |
+
},
|
| 2507 |
+
{
|
| 2508 |
+
"epoch": 0.96,
|
| 2509 |
+
"learning_rate": 3.9426493427611177e-07,
|
| 2510 |
+
"loss": 0.0,
|
| 2511 |
+
"step": 417
|
| 2512 |
+
},
|
| 2513 |
+
{
|
| 2514 |
+
"epoch": 0.96,
|
| 2515 |
+
"learning_rate": 3.492975587769898e-07,
|
| 2516 |
+
"loss": 0.0,
|
| 2517 |
+
"step": 418
|
| 2518 |
+
},
|
| 2519 |
+
{
|
| 2520 |
+
"epoch": 0.97,
|
| 2521 |
+
"learning_rate": 3.0704315523631953e-07,
|
| 2522 |
+
"loss": 0.0,
|
| 2523 |
+
"step": 419
|
| 2524 |
+
},
|
| 2525 |
+
{
|
| 2526 |
+
"epoch": 0.97,
|
| 2527 |
+
"learning_rate": 2.6750403248370636e-07,
|
| 2528 |
+
"loss": 0.0,
|
| 2529 |
+
"step": 420
|
| 2530 |
+
},
|
| 2531 |
+
{
|
| 2532 |
+
"epoch": 0.97,
|
| 2533 |
+
"learning_rate": 2.3068235098266944e-07,
|
| 2534 |
+
"loss": 0.0,
|
| 2535 |
+
"step": 421
|
| 2536 |
+
},
|
| 2537 |
+
{
|
| 2538 |
+
"epoch": 0.97,
|
| 2539 |
+
"learning_rate": 1.9658012271259185e-07,
|
| 2540 |
+
"loss": 0.0,
|
| 2541 |
+
"step": 422
|
| 2542 |
+
},
|
| 2543 |
+
{
|
| 2544 |
+
"epoch": 0.97,
|
| 2545 |
+
"learning_rate": 1.6519921105876412e-07,
|
| 2546 |
+
"loss": 0.0,
|
| 2547 |
+
"step": 423
|
| 2548 |
+
},
|
| 2549 |
+
{
|
| 2550 |
+
"epoch": 0.98,
|
| 2551 |
+
"learning_rate": 1.3654133071059893e-07,
|
| 2552 |
+
"loss": 0.0,
|
| 2553 |
+
"step": 424
|
| 2554 |
+
},
|
| 2555 |
+
{
|
| 2556 |
+
"epoch": 0.98,
|
| 2557 |
+
"learning_rate": 1.1060804756790055e-07,
|
| 2558 |
+
"loss": 0.0,
|
| 2559 |
+
"step": 425
|
| 2560 |
+
},
|
| 2561 |
+
{
|
| 2562 |
+
"epoch": 0.98,
|
| 2563 |
+
"learning_rate": 8.740077865534435e-08,
|
| 2564 |
+
"loss": 0.0,
|
| 2565 |
+
"step": 426
|
| 2566 |
+
},
|
| 2567 |
+
{
|
| 2568 |
+
"epoch": 0.98,
|
| 2569 |
+
"learning_rate": 6.692079204500545e-08,
|
| 2570 |
+
"loss": 0.0,
|
| 2571 |
+
"step": 427
|
| 2572 |
+
},
|
| 2573 |
+
{
|
| 2574 |
+
"epoch": 0.99,
|
| 2575 |
+
"learning_rate": 4.916920678710857e-08,
|
| 2576 |
+
"loss": 0.0499,
|
| 2577 |
+
"step": 428
|
| 2578 |
+
},
|
| 2579 |
+
{
|
| 2580 |
+
"epoch": 0.99,
|
| 2581 |
+
"learning_rate": 3.4146992848854695e-08,
|
| 2582 |
+
"loss": 0.0,
|
| 2583 |
+
"step": 429
|
| 2584 |
+
},
|
| 2585 |
+
{
|
| 2586 |
+
"epoch": 0.99,
|
| 2587 |
+
"learning_rate": 2.1854971061435747e-08,
|
| 2588 |
+
"loss": 0.0,
|
| 2589 |
+
"step": 430
|
| 2590 |
+
},
|
| 2591 |
+
{
|
| 2592 |
+
"epoch": 0.99,
|
| 2593 |
+
"learning_rate": 1.2293813075175963e-08,
|
| 2594 |
+
"loss": 0.0,
|
| 2595 |
+
"step": 431
|
| 2596 |
+
},
|
| 2597 |
+
{
|
| 2598 |
+
"epoch": 1.0,
|
| 2599 |
+
"learning_rate": 5.464041322833513e-09,
|
| 2600 |
+
"loss": 0.0658,
|
| 2601 |
+
"step": 432
|
| 2602 |
+
},
|
| 2603 |
+
{
|
| 2604 |
+
"epoch": 1.0,
|
| 2605 |
+
"learning_rate": 1.366028991062196e-09,
|
| 2606 |
+
"loss": 0.0473,
|
| 2607 |
+
"step": 433
|
| 2608 |
+
},
|
| 2609 |
+
{
|
| 2610 |
+
"epoch": 1.0,
|
| 2611 |
+
"learning_rate": 0.0,
|
| 2612 |
+
"loss": 0.0,
|
| 2613 |
+
"step": 434
|
| 2614 |
+
},
|
| 2615 |
+
{
|
| 2616 |
+
"epoch": 1.0,
|
| 2617 |
+
"step": 434,
|
| 2618 |
+
"total_flos": 510790631161856.0,
|
| 2619 |
+
"train_loss": 0.04482437436082502,
|
| 2620 |
+
"train_runtime": 12466.6764,
|
| 2621 |
+
"train_samples_per_second": 2.227,
|
| 2622 |
+
"train_steps_per_second": 0.035
|
| 2623 |
+
}
|
| 2624 |
+
],
|
| 2625 |
+
"logging_steps": 1.0,
|
| 2626 |
+
"max_steps": 434,
|
| 2627 |
+
"num_input_tokens_seen": 0,
|
| 2628 |
+
"num_train_epochs": 1,
|
| 2629 |
+
"save_steps": 50000,
|
| 2630 |
+
"total_flos": 510790631161856.0,
|
| 2631 |
+
"train_batch_size": 4,
|
| 2632 |
+
"trial_name": null,
|
| 2633 |
+
"trial_params": null
|
| 2634 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b88336da0247ffe15b807c920dc5ed2d3a368789ff9218020840f2c6673cbd58
|
| 3 |
+
size 7224
|
training_args.json
ADDED
|
@@ -0,0 +1,141 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"output_dir": "/home/cbn-gpu12/MedicalReport/finetuning/final_results/stage1_20250409/checkpoints",
|
| 3 |
+
"overwrite_output_dir": false,
|
| 4 |
+
"do_train": false,
|
| 5 |
+
"do_eval": false,
|
| 6 |
+
"do_predict": false,
|
| 7 |
+
"evaluation_strategy": "no",
|
| 8 |
+
"prediction_loss_only": false,
|
| 9 |
+
"per_device_train_batch_size": 4,
|
| 10 |
+
"per_device_eval_batch_size": 2,
|
| 11 |
+
"per_gpu_train_batch_size": null,
|
| 12 |
+
"per_gpu_eval_batch_size": null,
|
| 13 |
+
"gradient_accumulation_steps": 4,
|
| 14 |
+
"eval_accumulation_steps": null,
|
| 15 |
+
"eval_delay": 0,
|
| 16 |
+
"learning_rate": 0.0001,
|
| 17 |
+
"weight_decay": 0.01,
|
| 18 |
+
"adam_beta1": 0.9,
|
| 19 |
+
"adam_beta2": 0.999,
|
| 20 |
+
"adam_epsilon": 1e-08,
|
| 21 |
+
"max_grad_norm": 1.0,
|
| 22 |
+
"num_train_epochs": 1.0,
|
| 23 |
+
"max_steps": -1,
|
| 24 |
+
"lr_scheduler_type": "cosine",
|
| 25 |
+
"lr_scheduler_kwargs": {},
|
| 26 |
+
"warmup_ratio": 0.02,
|
| 27 |
+
"warmup_steps": 0,
|
| 28 |
+
"log_level": "passive",
|
| 29 |
+
"log_level_replica": "warning",
|
| 30 |
+
"log_on_each_node": true,
|
| 31 |
+
"logging_dir": "/home/cbn-gpu12/MedicalReport/finetuning/final_results/stage1_20250409/checkpoints/runs/Apr09_15-07-46_4b0e57b52168",
|
| 32 |
+
"logging_strategy": "steps",
|
| 33 |
+
"logging_first_step": false,
|
| 34 |
+
"logging_steps": 1.0,
|
| 35 |
+
"logging_nan_inf_filter": true,
|
| 36 |
+
"save_strategy": "steps",
|
| 37 |
+
"save_steps": 50000,
|
| 38 |
+
"save_total_limit": 1,
|
| 39 |
+
"save_safetensors": true,
|
| 40 |
+
"save_on_each_node": false,
|
| 41 |
+
"save_only_model": false,
|
| 42 |
+
"no_cuda": false,
|
| 43 |
+
"use_cpu": false,
|
| 44 |
+
"use_mps_device": false,
|
| 45 |
+
"seed": 42,
|
| 46 |
+
"data_seed": null,
|
| 47 |
+
"jit_mode_eval": false,
|
| 48 |
+
"use_ipex": false,
|
| 49 |
+
"bf16": false,
|
| 50 |
+
"fp16": true,
|
| 51 |
+
"fp16_opt_level": "O1",
|
| 52 |
+
"half_precision_backend": "auto",
|
| 53 |
+
"bf16_full_eval": false,
|
| 54 |
+
"fp16_full_eval": false,
|
| 55 |
+
"tf32": false,
|
| 56 |
+
"local_rank": 0,
|
| 57 |
+
"ddp_backend": null,
|
| 58 |
+
"tpu_num_cores": null,
|
| 59 |
+
"tpu_metrics_debug": false,
|
| 60 |
+
"debug": [],
|
| 61 |
+
"dataloader_drop_last": false,
|
| 62 |
+
"eval_steps": null,
|
| 63 |
+
"dataloader_num_workers": 4,
|
| 64 |
+
"past_index": -1,
|
| 65 |
+
"run_name": "llava-med-spect-persona-ffr-20250409",
|
| 66 |
+
"disable_tqdm": false,
|
| 67 |
+
"remove_unused_columns": false,
|
| 68 |
+
"label_names": null,
|
| 69 |
+
"load_best_model_at_end": false,
|
| 70 |
+
"metric_for_best_model": null,
|
| 71 |
+
"greater_is_better": null,
|
| 72 |
+
"ignore_data_skip": false,
|
| 73 |
+
"fsdp": [],
|
| 74 |
+
"fsdp_min_num_params": 0,
|
| 75 |
+
"fsdp_config": {
|
| 76 |
+
"min_num_params": 0,
|
| 77 |
+
"xla": false,
|
| 78 |
+
"xla_fsdp_grad_ckpt": false
|
| 79 |
+
},
|
| 80 |
+
"fsdp_transformer_layer_cls_to_wrap": null,
|
| 81 |
+
"deepspeed": "/home/cbn-gpu12/MedicalReport/LLaVA/scripts/zero3.json",
|
| 82 |
+
"label_smoothing_factor": 0.0,
|
| 83 |
+
"optim": "adamw_torch",
|
| 84 |
+
"optim_args": null,
|
| 85 |
+
"adafactor": false,
|
| 86 |
+
"group_by_length": false,
|
| 87 |
+
"length_column_name": "length",
|
| 88 |
+
"report_to": [
|
| 89 |
+
"wandb"
|
| 90 |
+
],
|
| 91 |
+
"ddp_find_unused_parameters": null,
|
| 92 |
+
"ddp_bucket_cap_mb": null,
|
| 93 |
+
"ddp_broadcast_buffers": null,
|
| 94 |
+
"dataloader_pin_memory": true,
|
| 95 |
+
"dataloader_persistent_workers": false,
|
| 96 |
+
"skip_memory_metrics": true,
|
| 97 |
+
"use_legacy_prediction_loop": false,
|
| 98 |
+
"push_to_hub": true,
|
| 99 |
+
"resume_from_checkpoint": null,
|
| 100 |
+
"hub_model_id": "sarahyo941/llava-med-spect-persona-ffr-20250409",
|
| 101 |
+
"hub_strategy": "every_save",
|
| 102 |
+
"hub_token": "<HUB_TOKEN>",
|
| 103 |
+
"hub_private_repo": false,
|
| 104 |
+
"hub_always_push": false,
|
| 105 |
+
"gradient_checkpointing": true,
|
| 106 |
+
"gradient_checkpointing_kwargs": null,
|
| 107 |
+
"include_inputs_for_metrics": false,
|
| 108 |
+
"fp16_backend": "auto",
|
| 109 |
+
"push_to_hub_model_id": null,
|
| 110 |
+
"push_to_hub_organization": null,
|
| 111 |
+
"push_to_hub_token": "<PUSH_TO_HUB_TOKEN>",
|
| 112 |
+
"mp_parameters": "",
|
| 113 |
+
"auto_find_batch_size": false,
|
| 114 |
+
"full_determinism": false,
|
| 115 |
+
"torchdynamo": null,
|
| 116 |
+
"ray_scope": "last",
|
| 117 |
+
"ddp_timeout": 1800,
|
| 118 |
+
"torch_compile": false,
|
| 119 |
+
"torch_compile_backend": null,
|
| 120 |
+
"torch_compile_mode": null,
|
| 121 |
+
"dispatch_batches": null,
|
| 122 |
+
"split_batches": false,
|
| 123 |
+
"include_tokens_per_second": false,
|
| 124 |
+
"include_num_input_tokens_seen": false,
|
| 125 |
+
"neftune_noise_alpha": null,
|
| 126 |
+
"cache_dir": null,
|
| 127 |
+
"freeze_mm_mlp_adapter": false,
|
| 128 |
+
"mpt_attn_impl": "triton",
|
| 129 |
+
"model_max_length": 2048,
|
| 130 |
+
"double_quant": true,
|
| 131 |
+
"quant_type": "nf4",
|
| 132 |
+
"bits": 16,
|
| 133 |
+
"lora_enable": true,
|
| 134 |
+
"lora_r": 16,
|
| 135 |
+
"lora_alpha": 32,
|
| 136 |
+
"lora_dropout": 0.05,
|
| 137 |
+
"lora_weight_path": "",
|
| 138 |
+
"lora_bias": "none",
|
| 139 |
+
"mm_projector_lr": 1e-05,
|
| 140 |
+
"group_by_modality_length": true
|
| 141 |
+
}
|