Upload folder using huggingface_hub
Browse files- README.md +16 -20
- checkpoint-497/config.json +28 -0
- checkpoint-497/generation_config.json +7 -0
- checkpoint-497/model.safetensors +3 -0
- checkpoint-497/optimizer.pt +3 -0
- checkpoint-497/rng_state.pth +3 -0
- checkpoint-497/scheduler.pt +3 -0
- checkpoint-497/trainer_state.json +3035 -0
- checkpoint-497/training_args.bin +3 -0
- pytorch_model.bin +1 -1
- runs/Feb08_19-02-05_ruche-gpu12.cluster/events.out.tfevents.1707415329.ruche-gpu12.cluster.7591.0 +3 -0
README.md
CHANGED
|
@@ -26,7 +26,7 @@ load_in_4bit: false
|
|
| 26 |
strict: false
|
| 27 |
|
| 28 |
datasets:
|
| 29 |
-
- path: manu/
|
| 30 |
split: train
|
| 31 |
type: completion
|
| 32 |
|
|
@@ -35,8 +35,8 @@ val_set_size: 0.05
|
|
| 35 |
output_dir: ./out
|
| 36 |
|
| 37 |
sequence_len: 2048
|
| 38 |
-
sample_packing:
|
| 39 |
-
pad_to_sequence_len:
|
| 40 |
|
| 41 |
adapter:
|
| 42 |
lora_model_dir:
|
|
@@ -52,9 +52,9 @@ wandb_watch:
|
|
| 52 |
wandb_name:
|
| 53 |
wandb_log_model:
|
| 54 |
|
| 55 |
-
gradient_accumulation_steps:
|
| 56 |
-
micro_batch_size:
|
| 57 |
-
num_epochs:
|
| 58 |
optimizer: adamw_bnb_8bit
|
| 59 |
lr_scheduler: cosine
|
| 60 |
learning_rate: 0.0002
|
|
@@ -96,7 +96,7 @@ special_tokens:
|
|
| 96 |
|
| 97 |
This model is a fine-tuned version of [croissantllm/CroissantLLMBase](https://huggingface.co/croissantllm/CroissantLLMBase) on the None dataset.
|
| 98 |
It achieves the following results on the evaluation set:
|
| 99 |
-
- Loss:
|
| 100 |
|
| 101 |
## Model description
|
| 102 |
|
|
@@ -116,28 +116,24 @@ More information needed
|
|
| 116 |
|
| 117 |
The following hyperparameters were used during training:
|
| 118 |
- learning_rate: 0.0002
|
| 119 |
-
- train_batch_size:
|
| 120 |
-
- eval_batch_size:
|
| 121 |
- seed: 42
|
| 122 |
-
- gradient_accumulation_steps:
|
| 123 |
-
- total_train_batch_size:
|
| 124 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 125 |
- lr_scheduler_type: cosine
|
| 126 |
- lr_scheduler_warmup_steps: 50
|
| 127 |
-
- num_epochs:
|
| 128 |
|
| 129 |
### Training results
|
| 130 |
|
| 131 |
| Training Loss | Epoch | Step | Validation Loss |
|
| 132 |
|:-------------:|:-----:|:----:|:---------------:|
|
| 133 |
-
| 2.
|
| 134 |
-
| 2.
|
| 135 |
-
| 1.
|
| 136 |
-
| 1.
|
| 137 |
-
| 1.3767 | 1.0 | 276 | 1.3323 |
|
| 138 |
-
| 0.7872 | 1.23 | 345 | 1.0583 |
|
| 139 |
-
| 0.5873 | 1.48 | 414 | 0.8251 |
|
| 140 |
-
| 0.5154 | 1.73 | 483 | 0.7378 |
|
| 141 |
|
| 142 |
|
| 143 |
### Framework versions
|
|
|
|
| 26 |
strict: false
|
| 27 |
|
| 28 |
datasets:
|
| 29 |
+
- path: manu/mmlu_auxiliary_train_formatted_3
|
| 30 |
split: train
|
| 31 |
type: completion
|
| 32 |
|
|
|
|
| 35 |
output_dir: ./out
|
| 36 |
|
| 37 |
sequence_len: 2048
|
| 38 |
+
sample_packing: false
|
| 39 |
+
pad_to_sequence_len: false
|
| 40 |
|
| 41 |
adapter:
|
| 42 |
lora_model_dir:
|
|
|
|
| 52 |
wandb_name:
|
| 53 |
wandb_log_model:
|
| 54 |
|
| 55 |
+
gradient_accumulation_steps: 2
|
| 56 |
+
micro_batch_size: 24
|
| 57 |
+
num_epochs: 1
|
| 58 |
optimizer: adamw_bnb_8bit
|
| 59 |
lr_scheduler: cosine
|
| 60 |
learning_rate: 0.0002
|
|
|
|
| 96 |
|
| 97 |
This model is a fine-tuned version of [croissantllm/CroissantLLMBase](https://huggingface.co/croissantllm/CroissantLLMBase) on the None dataset.
|
| 98 |
It achieves the following results on the evaluation set:
|
| 99 |
+
- Loss: 1.3916
|
| 100 |
|
| 101 |
## Model description
|
| 102 |
|
|
|
|
| 116 |
|
| 117 |
The following hyperparameters were used during training:
|
| 118 |
- learning_rate: 0.0002
|
| 119 |
+
- train_batch_size: 24
|
| 120 |
+
- eval_batch_size: 24
|
| 121 |
- seed: 42
|
| 122 |
+
- gradient_accumulation_steps: 2
|
| 123 |
+
- total_train_batch_size: 48
|
| 124 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 125 |
- lr_scheduler_type: cosine
|
| 126 |
- lr_scheduler_warmup_steps: 50
|
| 127 |
+
- num_epochs: 1
|
| 128 |
|
| 129 |
### Training results
|
| 130 |
|
| 131 |
| Training Loss | Epoch | Step | Validation Loss |
|
| 132 |
|:-------------:|:-----:|:----:|:---------------:|
|
| 133 |
+
| 2.5221 | 0.0 | 1 | 2.5482 |
|
| 134 |
+
| 2.2198 | 0.25 | 125 | 2.3040 |
|
| 135 |
+
| 1.8488 | 0.5 | 250 | 1.8220 |
|
| 136 |
+
| 1.383 | 0.75 | 375 | 1.3916 |
|
|
|
|
|
|
|
|
|
|
|
|
|
| 137 |
|
| 138 |
|
| 139 |
### Framework versions
|
checkpoint-497/config.json
ADDED
|
@@ -0,0 +1,28 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "croissantllm/CroissantLLMBase",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"LlamaForCausalLM"
|
| 5 |
+
],
|
| 6 |
+
"attention_bias": false,
|
| 7 |
+
"attention_dropout": 0.0,
|
| 8 |
+
"bos_token_id": 1,
|
| 9 |
+
"eos_token_id": 2,
|
| 10 |
+
"hidden_act": "silu",
|
| 11 |
+
"hidden_size": 2048,
|
| 12 |
+
"initializer_range": 0.02,
|
| 13 |
+
"intermediate_size": 5504,
|
| 14 |
+
"max_position_embeddings": 2048,
|
| 15 |
+
"model_type": "llama",
|
| 16 |
+
"num_attention_heads": 16,
|
| 17 |
+
"num_hidden_layers": 24,
|
| 18 |
+
"num_key_value_heads": 16,
|
| 19 |
+
"pretraining_tp": 1,
|
| 20 |
+
"rms_norm_eps": 1e-05,
|
| 21 |
+
"rope_scaling": null,
|
| 22 |
+
"rope_theta": 10000.0,
|
| 23 |
+
"tie_word_embeddings": false,
|
| 24 |
+
"torch_dtype": "bfloat16",
|
| 25 |
+
"transformers_version": "4.38.0.dev0",
|
| 26 |
+
"use_cache": false,
|
| 27 |
+
"vocab_size": 32000
|
| 28 |
+
}
|
checkpoint-497/generation_config.json
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"bos_token_id": 1,
|
| 4 |
+
"do_sample": true,
|
| 5 |
+
"eos_token_id": 2,
|
| 6 |
+
"transformers_version": "4.38.0.dev0"
|
| 7 |
+
}
|
checkpoint-497/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a40e2e6aaf6ce330f8a6086126be8e2de420b4027e9dfa8bfda4375d99f2d859
|
| 3 |
+
size 2690869336
|
checkpoint-497/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b348baf6696f1a572ed746ed0121d53ded698752c88031b0442e619aef506df0
|
| 3 |
+
size 2696906170
|
checkpoint-497/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c062f7f375beded48b5337f5a3f3a5cb38807fa3e85dbf3e294c0ab6b627bfc2
|
| 3 |
+
size 14244
|
checkpoint-497/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7d0a3c0e024c32aabc66fb0ac86f212ddd7129495621b7e5eb772663d347fb60
|
| 3 |
+
size 1064
|
checkpoint-497/trainer_state.json
ADDED
|
@@ -0,0 +1,3035 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.9989949748743718,
|
| 5 |
+
"eval_steps": 125,
|
| 6 |
+
"global_step": 497,
|
| 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": 4.000000000000001e-06,
|
| 14 |
+
"loss": 2.5221,
|
| 15 |
+
"step": 1
|
| 16 |
+
},
|
| 17 |
+
{
|
| 18 |
+
"epoch": 0.0,
|
| 19 |
+
"eval_loss": 2.548204183578491,
|
| 20 |
+
"eval_runtime": 51.7546,
|
| 21 |
+
"eval_samples_per_second": 24.288,
|
| 22 |
+
"eval_steps_per_second": 1.024,
|
| 23 |
+
"step": 1
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.0,
|
| 27 |
+
"learning_rate": 8.000000000000001e-06,
|
| 28 |
+
"loss": 2.5234,
|
| 29 |
+
"step": 2
|
| 30 |
+
},
|
| 31 |
+
{
|
| 32 |
+
"epoch": 0.01,
|
| 33 |
+
"learning_rate": 1.2e-05,
|
| 34 |
+
"loss": 2.55,
|
| 35 |
+
"step": 3
|
| 36 |
+
},
|
| 37 |
+
{
|
| 38 |
+
"epoch": 0.01,
|
| 39 |
+
"learning_rate": 1.6000000000000003e-05,
|
| 40 |
+
"loss": 2.5209,
|
| 41 |
+
"step": 4
|
| 42 |
+
},
|
| 43 |
+
{
|
| 44 |
+
"epoch": 0.01,
|
| 45 |
+
"learning_rate": 2e-05,
|
| 46 |
+
"loss": 2.5177,
|
| 47 |
+
"step": 5
|
| 48 |
+
},
|
| 49 |
+
{
|
| 50 |
+
"epoch": 0.01,
|
| 51 |
+
"learning_rate": 2.4e-05,
|
| 52 |
+
"loss": 2.4932,
|
| 53 |
+
"step": 6
|
| 54 |
+
},
|
| 55 |
+
{
|
| 56 |
+
"epoch": 0.01,
|
| 57 |
+
"learning_rate": 2.8000000000000003e-05,
|
| 58 |
+
"loss": 2.495,
|
| 59 |
+
"step": 7
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 0.02,
|
| 63 |
+
"learning_rate": 3.2000000000000005e-05,
|
| 64 |
+
"loss": 2.4618,
|
| 65 |
+
"step": 8
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.02,
|
| 69 |
+
"learning_rate": 3.6e-05,
|
| 70 |
+
"loss": 2.4603,
|
| 71 |
+
"step": 9
|
| 72 |
+
},
|
| 73 |
+
{
|
| 74 |
+
"epoch": 0.02,
|
| 75 |
+
"learning_rate": 4e-05,
|
| 76 |
+
"loss": 2.4364,
|
| 77 |
+
"step": 10
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 0.02,
|
| 81 |
+
"learning_rate": 4.4000000000000006e-05,
|
| 82 |
+
"loss": 2.4445,
|
| 83 |
+
"step": 11
|
| 84 |
+
},
|
| 85 |
+
{
|
| 86 |
+
"epoch": 0.02,
|
| 87 |
+
"learning_rate": 4.8e-05,
|
| 88 |
+
"loss": 2.4126,
|
| 89 |
+
"step": 12
|
| 90 |
+
},
|
| 91 |
+
{
|
| 92 |
+
"epoch": 0.03,
|
| 93 |
+
"learning_rate": 5.2000000000000004e-05,
|
| 94 |
+
"loss": 2.4573,
|
| 95 |
+
"step": 13
|
| 96 |
+
},
|
| 97 |
+
{
|
| 98 |
+
"epoch": 0.03,
|
| 99 |
+
"learning_rate": 5.6000000000000006e-05,
|
| 100 |
+
"loss": 2.4604,
|
| 101 |
+
"step": 14
|
| 102 |
+
},
|
| 103 |
+
{
|
| 104 |
+
"epoch": 0.03,
|
| 105 |
+
"learning_rate": 6e-05,
|
| 106 |
+
"loss": 2.4551,
|
| 107 |
+
"step": 15
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.03,
|
| 111 |
+
"learning_rate": 6.400000000000001e-05,
|
| 112 |
+
"loss": 2.4235,
|
| 113 |
+
"step": 16
|
| 114 |
+
},
|
| 115 |
+
{
|
| 116 |
+
"epoch": 0.03,
|
| 117 |
+
"learning_rate": 6.800000000000001e-05,
|
| 118 |
+
"loss": 2.442,
|
| 119 |
+
"step": 17
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 0.04,
|
| 123 |
+
"learning_rate": 7.2e-05,
|
| 124 |
+
"loss": 2.4189,
|
| 125 |
+
"step": 18
|
| 126 |
+
},
|
| 127 |
+
{
|
| 128 |
+
"epoch": 0.04,
|
| 129 |
+
"learning_rate": 7.6e-05,
|
| 130 |
+
"loss": 2.4027,
|
| 131 |
+
"step": 19
|
| 132 |
+
},
|
| 133 |
+
{
|
| 134 |
+
"epoch": 0.04,
|
| 135 |
+
"learning_rate": 8e-05,
|
| 136 |
+
"loss": 2.4114,
|
| 137 |
+
"step": 20
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 0.04,
|
| 141 |
+
"learning_rate": 8.4e-05,
|
| 142 |
+
"loss": 2.4136,
|
| 143 |
+
"step": 21
|
| 144 |
+
},
|
| 145 |
+
{
|
| 146 |
+
"epoch": 0.04,
|
| 147 |
+
"learning_rate": 8.800000000000001e-05,
|
| 148 |
+
"loss": 2.4136,
|
| 149 |
+
"step": 22
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.05,
|
| 153 |
+
"learning_rate": 9.200000000000001e-05,
|
| 154 |
+
"loss": 2.4229,
|
| 155 |
+
"step": 23
|
| 156 |
+
},
|
| 157 |
+
{
|
| 158 |
+
"epoch": 0.05,
|
| 159 |
+
"learning_rate": 9.6e-05,
|
| 160 |
+
"loss": 2.4245,
|
| 161 |
+
"step": 24
|
| 162 |
+
},
|
| 163 |
+
{
|
| 164 |
+
"epoch": 0.05,
|
| 165 |
+
"learning_rate": 0.0001,
|
| 166 |
+
"loss": 2.3903,
|
| 167 |
+
"step": 25
|
| 168 |
+
},
|
| 169 |
+
{
|
| 170 |
+
"epoch": 0.05,
|
| 171 |
+
"learning_rate": 0.00010400000000000001,
|
| 172 |
+
"loss": 2.4154,
|
| 173 |
+
"step": 26
|
| 174 |
+
},
|
| 175 |
+
{
|
| 176 |
+
"epoch": 0.05,
|
| 177 |
+
"learning_rate": 0.00010800000000000001,
|
| 178 |
+
"loss": 2.46,
|
| 179 |
+
"step": 27
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 0.06,
|
| 183 |
+
"learning_rate": 0.00011200000000000001,
|
| 184 |
+
"loss": 2.4382,
|
| 185 |
+
"step": 28
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 0.06,
|
| 189 |
+
"learning_rate": 0.000116,
|
| 190 |
+
"loss": 2.3889,
|
| 191 |
+
"step": 29
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.06,
|
| 195 |
+
"learning_rate": 0.00012,
|
| 196 |
+
"loss": 2.4225,
|
| 197 |
+
"step": 30
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 0.06,
|
| 201 |
+
"learning_rate": 0.000124,
|
| 202 |
+
"loss": 2.4394,
|
| 203 |
+
"step": 31
|
| 204 |
+
},
|
| 205 |
+
{
|
| 206 |
+
"epoch": 0.06,
|
| 207 |
+
"learning_rate": 0.00012800000000000002,
|
| 208 |
+
"loss": 2.4058,
|
| 209 |
+
"step": 32
|
| 210 |
+
},
|
| 211 |
+
{
|
| 212 |
+
"epoch": 0.07,
|
| 213 |
+
"learning_rate": 0.000132,
|
| 214 |
+
"loss": 2.448,
|
| 215 |
+
"step": 33
|
| 216 |
+
},
|
| 217 |
+
{
|
| 218 |
+
"epoch": 0.07,
|
| 219 |
+
"learning_rate": 0.00013600000000000003,
|
| 220 |
+
"loss": 2.4017,
|
| 221 |
+
"step": 34
|
| 222 |
+
},
|
| 223 |
+
{
|
| 224 |
+
"epoch": 0.07,
|
| 225 |
+
"learning_rate": 0.00014,
|
| 226 |
+
"loss": 2.4179,
|
| 227 |
+
"step": 35
|
| 228 |
+
},
|
| 229 |
+
{
|
| 230 |
+
"epoch": 0.07,
|
| 231 |
+
"learning_rate": 0.000144,
|
| 232 |
+
"loss": 2.4142,
|
| 233 |
+
"step": 36
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.07,
|
| 237 |
+
"learning_rate": 0.000148,
|
| 238 |
+
"loss": 2.419,
|
| 239 |
+
"step": 37
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 0.08,
|
| 243 |
+
"learning_rate": 0.000152,
|
| 244 |
+
"loss": 2.3895,
|
| 245 |
+
"step": 38
|
| 246 |
+
},
|
| 247 |
+
{
|
| 248 |
+
"epoch": 0.08,
|
| 249 |
+
"learning_rate": 0.00015600000000000002,
|
| 250 |
+
"loss": 2.3415,
|
| 251 |
+
"step": 39
|
| 252 |
+
},
|
| 253 |
+
{
|
| 254 |
+
"epoch": 0.08,
|
| 255 |
+
"learning_rate": 0.00016,
|
| 256 |
+
"loss": 2.4038,
|
| 257 |
+
"step": 40
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 0.08,
|
| 261 |
+
"learning_rate": 0.000164,
|
| 262 |
+
"loss": 2.415,
|
| 263 |
+
"step": 41
|
| 264 |
+
},
|
| 265 |
+
{
|
| 266 |
+
"epoch": 0.08,
|
| 267 |
+
"learning_rate": 0.000168,
|
| 268 |
+
"loss": 2.3996,
|
| 269 |
+
"step": 42
|
| 270 |
+
},
|
| 271 |
+
{
|
| 272 |
+
"epoch": 0.09,
|
| 273 |
+
"learning_rate": 0.000172,
|
| 274 |
+
"loss": 2.4102,
|
| 275 |
+
"step": 43
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.09,
|
| 279 |
+
"learning_rate": 0.00017600000000000002,
|
| 280 |
+
"loss": 2.4241,
|
| 281 |
+
"step": 44
|
| 282 |
+
},
|
| 283 |
+
{
|
| 284 |
+
"epoch": 0.09,
|
| 285 |
+
"learning_rate": 0.00018,
|
| 286 |
+
"loss": 2.4232,
|
| 287 |
+
"step": 45
|
| 288 |
+
},
|
| 289 |
+
{
|
| 290 |
+
"epoch": 0.09,
|
| 291 |
+
"learning_rate": 0.00018400000000000003,
|
| 292 |
+
"loss": 2.4156,
|
| 293 |
+
"step": 46
|
| 294 |
+
},
|
| 295 |
+
{
|
| 296 |
+
"epoch": 0.09,
|
| 297 |
+
"learning_rate": 0.000188,
|
| 298 |
+
"loss": 2.4106,
|
| 299 |
+
"step": 47
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 0.1,
|
| 303 |
+
"learning_rate": 0.000192,
|
| 304 |
+
"loss": 2.4124,
|
| 305 |
+
"step": 48
|
| 306 |
+
},
|
| 307 |
+
{
|
| 308 |
+
"epoch": 0.1,
|
| 309 |
+
"learning_rate": 0.000196,
|
| 310 |
+
"loss": 2.4515,
|
| 311 |
+
"step": 49
|
| 312 |
+
},
|
| 313 |
+
{
|
| 314 |
+
"epoch": 0.1,
|
| 315 |
+
"learning_rate": 0.0002,
|
| 316 |
+
"loss": 2.3911,
|
| 317 |
+
"step": 50
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.1,
|
| 321 |
+
"learning_rate": 0.00019999753025044538,
|
| 322 |
+
"loss": 2.437,
|
| 323 |
+
"step": 51
|
| 324 |
+
},
|
| 325 |
+
{
|
| 326 |
+
"epoch": 0.1,
|
| 327 |
+
"learning_rate": 0.00019999012112377473,
|
| 328 |
+
"loss": 2.4317,
|
| 329 |
+
"step": 52
|
| 330 |
+
},
|
| 331 |
+
{
|
| 332 |
+
"epoch": 0.11,
|
| 333 |
+
"learning_rate": 0.0001999777729859618,
|
| 334 |
+
"loss": 2.4115,
|
| 335 |
+
"step": 53
|
| 336 |
+
},
|
| 337 |
+
{
|
| 338 |
+
"epoch": 0.11,
|
| 339 |
+
"learning_rate": 0.00019996048644694282,
|
| 340 |
+
"loss": 2.482,
|
| 341 |
+
"step": 54
|
| 342 |
+
},
|
| 343 |
+
{
|
| 344 |
+
"epoch": 0.11,
|
| 345 |
+
"learning_rate": 0.00019993826236058612,
|
| 346 |
+
"loss": 2.4374,
|
| 347 |
+
"step": 55
|
| 348 |
+
},
|
| 349 |
+
{
|
| 350 |
+
"epoch": 0.11,
|
| 351 |
+
"learning_rate": 0.00019991110182465032,
|
| 352 |
+
"loss": 2.4609,
|
| 353 |
+
"step": 56
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"epoch": 0.11,
|
| 357 |
+
"learning_rate": 0.0001998790061807298,
|
| 358 |
+
"loss": 2.4661,
|
| 359 |
+
"step": 57
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.12,
|
| 363 |
+
"learning_rate": 0.0001998419770141887,
|
| 364 |
+
"loss": 2.4656,
|
| 365 |
+
"step": 58
|
| 366 |
+
},
|
| 367 |
+
{
|
| 368 |
+
"epoch": 0.12,
|
| 369 |
+
"learning_rate": 0.00019980001615408228,
|
| 370 |
+
"loss": 2.4659,
|
| 371 |
+
"step": 59
|
| 372 |
+
},
|
| 373 |
+
{
|
| 374 |
+
"epoch": 0.12,
|
| 375 |
+
"learning_rate": 0.00019975312567306687,
|
| 376 |
+
"loss": 2.4142,
|
| 377 |
+
"step": 60
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 0.12,
|
| 381 |
+
"learning_rate": 0.00019970130788729738,
|
| 382 |
+
"loss": 2.4815,
|
| 383 |
+
"step": 61
|
| 384 |
+
},
|
| 385 |
+
{
|
| 386 |
+
"epoch": 0.12,
|
| 387 |
+
"learning_rate": 0.00019964456535631286,
|
| 388 |
+
"loss": 2.4746,
|
| 389 |
+
"step": 62
|
| 390 |
+
},
|
| 391 |
+
{
|
| 392 |
+
"epoch": 0.13,
|
| 393 |
+
"learning_rate": 0.00019958290088291011,
|
| 394 |
+
"loss": 2.4358,
|
| 395 |
+
"step": 63
|
| 396 |
+
},
|
| 397 |
+
{
|
| 398 |
+
"epoch": 0.13,
|
| 399 |
+
"learning_rate": 0.00019951631751300531,
|
| 400 |
+
"loss": 2.4215,
|
| 401 |
+
"step": 64
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.13,
|
| 405 |
+
"learning_rate": 0.00019944481853548335,
|
| 406 |
+
"loss": 2.4716,
|
| 407 |
+
"step": 65
|
| 408 |
+
},
|
| 409 |
+
{
|
| 410 |
+
"epoch": 0.13,
|
| 411 |
+
"learning_rate": 0.00019936840748203561,
|
| 412 |
+
"loss": 2.3693,
|
| 413 |
+
"step": 66
|
| 414 |
+
},
|
| 415 |
+
{
|
| 416 |
+
"epoch": 0.13,
|
| 417 |
+
"learning_rate": 0.00019928708812698545,
|
| 418 |
+
"loss": 2.4516,
|
| 419 |
+
"step": 67
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 0.14,
|
| 423 |
+
"learning_rate": 0.0001992008644871016,
|
| 424 |
+
"loss": 2.4226,
|
| 425 |
+
"step": 68
|
| 426 |
+
},
|
| 427 |
+
{
|
| 428 |
+
"epoch": 0.14,
|
| 429 |
+
"learning_rate": 0.00019910974082140004,
|
| 430 |
+
"loss": 2.4769,
|
| 431 |
+
"step": 69
|
| 432 |
+
},
|
| 433 |
+
{
|
| 434 |
+
"epoch": 0.14,
|
| 435 |
+
"learning_rate": 0.0001990137216309334,
|
| 436 |
+
"loss": 2.4157,
|
| 437 |
+
"step": 70
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 0.14,
|
| 441 |
+
"learning_rate": 0.00019891281165856873,
|
| 442 |
+
"loss": 2.5004,
|
| 443 |
+
"step": 71
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.14,
|
| 447 |
+
"learning_rate": 0.00019880701588875327,
|
| 448 |
+
"loss": 2.4319,
|
| 449 |
+
"step": 72
|
| 450 |
+
},
|
| 451 |
+
{
|
| 452 |
+
"epoch": 0.15,
|
| 453 |
+
"learning_rate": 0.00019869633954726807,
|
| 454 |
+
"loss": 2.4563,
|
| 455 |
+
"step": 73
|
| 456 |
+
},
|
| 457 |
+
{
|
| 458 |
+
"epoch": 0.15,
|
| 459 |
+
"learning_rate": 0.00019858078810097002,
|
| 460 |
+
"loss": 2.4156,
|
| 461 |
+
"step": 74
|
| 462 |
+
},
|
| 463 |
+
{
|
| 464 |
+
"epoch": 0.15,
|
| 465 |
+
"learning_rate": 0.00019846036725752186,
|
| 466 |
+
"loss": 2.409,
|
| 467 |
+
"step": 75
|
| 468 |
+
},
|
| 469 |
+
{
|
| 470 |
+
"epoch": 0.15,
|
| 471 |
+
"learning_rate": 0.00019833508296511002,
|
| 472 |
+
"loss": 2.4533,
|
| 473 |
+
"step": 76
|
| 474 |
+
},
|
| 475 |
+
{
|
| 476 |
+
"epoch": 0.15,
|
| 477 |
+
"learning_rate": 0.00019820494141215104,
|
| 478 |
+
"loss": 2.4374,
|
| 479 |
+
"step": 77
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 0.16,
|
| 483 |
+
"learning_rate": 0.00019806994902698573,
|
| 484 |
+
"loss": 2.3888,
|
| 485 |
+
"step": 78
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 0.16,
|
| 489 |
+
"learning_rate": 0.00019793011247756174,
|
| 490 |
+
"loss": 2.4201,
|
| 491 |
+
"step": 79
|
| 492 |
+
},
|
| 493 |
+
{
|
| 494 |
+
"epoch": 0.16,
|
| 495 |
+
"learning_rate": 0.00019778543867110426,
|
| 496 |
+
"loss": 2.4034,
|
| 497 |
+
"step": 80
|
| 498 |
+
},
|
| 499 |
+
{
|
| 500 |
+
"epoch": 0.16,
|
| 501 |
+
"learning_rate": 0.00019763593475377462,
|
| 502 |
+
"loss": 2.442,
|
| 503 |
+
"step": 81
|
| 504 |
+
},
|
| 505 |
+
{
|
| 506 |
+
"epoch": 0.16,
|
| 507 |
+
"learning_rate": 0.00019748160811031747,
|
| 508 |
+
"loss": 2.3813,
|
| 509 |
+
"step": 82
|
| 510 |
+
},
|
| 511 |
+
{
|
| 512 |
+
"epoch": 0.17,
|
| 513 |
+
"learning_rate": 0.00019732246636369605,
|
| 514 |
+
"loss": 2.3356,
|
| 515 |
+
"step": 83
|
| 516 |
+
},
|
| 517 |
+
{
|
| 518 |
+
"epoch": 0.17,
|
| 519 |
+
"learning_rate": 0.00019715851737471546,
|
| 520 |
+
"loss": 2.3447,
|
| 521 |
+
"step": 84
|
| 522 |
+
},
|
| 523 |
+
{
|
| 524 |
+
"epoch": 0.17,
|
| 525 |
+
"learning_rate": 0.00019698976924163456,
|
| 526 |
+
"loss": 2.3692,
|
| 527 |
+
"step": 85
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 0.17,
|
| 531 |
+
"learning_rate": 0.00019681623029976588,
|
| 532 |
+
"loss": 2.3599,
|
| 533 |
+
"step": 86
|
| 534 |
+
},
|
| 535 |
+
{
|
| 536 |
+
"epoch": 0.17,
|
| 537 |
+
"learning_rate": 0.00019663790912106393,
|
| 538 |
+
"loss": 2.4325,
|
| 539 |
+
"step": 87
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"epoch": 0.18,
|
| 543 |
+
"learning_rate": 0.00019645481451370172,
|
| 544 |
+
"loss": 2.4405,
|
| 545 |
+
"step": 88
|
| 546 |
+
},
|
| 547 |
+
{
|
| 548 |
+
"epoch": 0.18,
|
| 549 |
+
"learning_rate": 0.00019626695552163578,
|
| 550 |
+
"loss": 2.397,
|
| 551 |
+
"step": 89
|
| 552 |
+
},
|
| 553 |
+
{
|
| 554 |
+
"epoch": 0.18,
|
| 555 |
+
"learning_rate": 0.0001960743414241593,
|
| 556 |
+
"loss": 2.3989,
|
| 557 |
+
"step": 90
|
| 558 |
+
},
|
| 559 |
+
{
|
| 560 |
+
"epoch": 0.18,
|
| 561 |
+
"learning_rate": 0.00019587698173544396,
|
| 562 |
+
"loss": 2.3333,
|
| 563 |
+
"step": 91
|
| 564 |
+
},
|
| 565 |
+
{
|
| 566 |
+
"epoch": 0.18,
|
| 567 |
+
"learning_rate": 0.00019567488620406983,
|
| 568 |
+
"loss": 2.3499,
|
| 569 |
+
"step": 92
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 0.19,
|
| 573 |
+
"learning_rate": 0.0001954680648125438,
|
| 574 |
+
"loss": 2.3654,
|
| 575 |
+
"step": 93
|
| 576 |
+
},
|
| 577 |
+
{
|
| 578 |
+
"epoch": 0.19,
|
| 579 |
+
"learning_rate": 0.00019525652777680676,
|
| 580 |
+
"loss": 2.3604,
|
| 581 |
+
"step": 94
|
| 582 |
+
},
|
| 583 |
+
{
|
| 584 |
+
"epoch": 0.19,
|
| 585 |
+
"learning_rate": 0.00019504028554572864,
|
| 586 |
+
"loss": 2.368,
|
| 587 |
+
"step": 95
|
| 588 |
+
},
|
| 589 |
+
{
|
| 590 |
+
"epoch": 0.19,
|
| 591 |
+
"learning_rate": 0.00019481934880059257,
|
| 592 |
+
"loss": 2.3225,
|
| 593 |
+
"step": 96
|
| 594 |
+
},
|
| 595 |
+
{
|
| 596 |
+
"epoch": 0.19,
|
| 597 |
+
"learning_rate": 0.00019459372845456705,
|
| 598 |
+
"loss": 2.3822,
|
| 599 |
+
"step": 97
|
| 600 |
+
},
|
| 601 |
+
{
|
| 602 |
+
"epoch": 0.2,
|
| 603 |
+
"learning_rate": 0.00019436343565216711,
|
| 604 |
+
"loss": 2.3697,
|
| 605 |
+
"step": 98
|
| 606 |
+
},
|
| 607 |
+
{
|
| 608 |
+
"epoch": 0.2,
|
| 609 |
+
"learning_rate": 0.00019412848176870363,
|
| 610 |
+
"loss": 2.349,
|
| 611 |
+
"step": 99
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 0.2,
|
| 615 |
+
"learning_rate": 0.00019388887840972164,
|
| 616 |
+
"loss": 2.3015,
|
| 617 |
+
"step": 100
|
| 618 |
+
},
|
| 619 |
+
{
|
| 620 |
+
"epoch": 0.2,
|
| 621 |
+
"learning_rate": 0.00019364463741042694,
|
| 622 |
+
"loss": 2.4336,
|
| 623 |
+
"step": 101
|
| 624 |
+
},
|
| 625 |
+
{
|
| 626 |
+
"epoch": 0.21,
|
| 627 |
+
"learning_rate": 0.00019339577083510144,
|
| 628 |
+
"loss": 2.2816,
|
| 629 |
+
"step": 102
|
| 630 |
+
},
|
| 631 |
+
{
|
| 632 |
+
"epoch": 0.21,
|
| 633 |
+
"learning_rate": 0.00019314229097650742,
|
| 634 |
+
"loss": 2.3033,
|
| 635 |
+
"step": 103
|
| 636 |
+
},
|
| 637 |
+
{
|
| 638 |
+
"epoch": 0.21,
|
| 639 |
+
"learning_rate": 0.00019288421035528028,
|
| 640 |
+
"loss": 2.372,
|
| 641 |
+
"step": 104
|
| 642 |
+
},
|
| 643 |
+
{
|
| 644 |
+
"epoch": 0.21,
|
| 645 |
+
"learning_rate": 0.00019262154171931,
|
| 646 |
+
"loss": 2.2949,
|
| 647 |
+
"step": 105
|
| 648 |
+
},
|
| 649 |
+
{
|
| 650 |
+
"epoch": 0.21,
|
| 651 |
+
"learning_rate": 0.0001923542980431115,
|
| 652 |
+
"loss": 2.3972,
|
| 653 |
+
"step": 106
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 0.22,
|
| 657 |
+
"learning_rate": 0.0001920824925271838,
|
| 658 |
+
"loss": 2.3624,
|
| 659 |
+
"step": 107
|
| 660 |
+
},
|
| 661 |
+
{
|
| 662 |
+
"epoch": 0.22,
|
| 663 |
+
"learning_rate": 0.00019180613859735791,
|
| 664 |
+
"loss": 2.4214,
|
| 665 |
+
"step": 108
|
| 666 |
+
},
|
| 667 |
+
{
|
| 668 |
+
"epoch": 0.22,
|
| 669 |
+
"learning_rate": 0.0001915252499041338,
|
| 670 |
+
"loss": 2.2428,
|
| 671 |
+
"step": 109
|
| 672 |
+
},
|
| 673 |
+
{
|
| 674 |
+
"epoch": 0.22,
|
| 675 |
+
"learning_rate": 0.00019123984032200586,
|
| 676 |
+
"loss": 2.3559,
|
| 677 |
+
"step": 110
|
| 678 |
+
},
|
| 679 |
+
{
|
| 680 |
+
"epoch": 0.22,
|
| 681 |
+
"learning_rate": 0.00019094992394877794,
|
| 682 |
+
"loss": 2.3033,
|
| 683 |
+
"step": 111
|
| 684 |
+
},
|
| 685 |
+
{
|
| 686 |
+
"epoch": 0.23,
|
| 687 |
+
"learning_rate": 0.0001906555151048667,
|
| 688 |
+
"loss": 2.3734,
|
| 689 |
+
"step": 112
|
| 690 |
+
},
|
| 691 |
+
{
|
| 692 |
+
"epoch": 0.23,
|
| 693 |
+
"learning_rate": 0.00019035662833259432,
|
| 694 |
+
"loss": 2.2408,
|
| 695 |
+
"step": 113
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"epoch": 0.23,
|
| 699 |
+
"learning_rate": 0.0001900532783954703,
|
| 700 |
+
"loss": 2.3528,
|
| 701 |
+
"step": 114
|
| 702 |
+
},
|
| 703 |
+
{
|
| 704 |
+
"epoch": 0.23,
|
| 705 |
+
"learning_rate": 0.0001897454802774621,
|
| 706 |
+
"loss": 2.2373,
|
| 707 |
+
"step": 115
|
| 708 |
+
},
|
| 709 |
+
{
|
| 710 |
+
"epoch": 0.23,
|
| 711 |
+
"learning_rate": 0.00018943324918225494,
|
| 712 |
+
"loss": 2.3138,
|
| 713 |
+
"step": 116
|
| 714 |
+
},
|
| 715 |
+
{
|
| 716 |
+
"epoch": 0.24,
|
| 717 |
+
"learning_rate": 0.00018911660053250103,
|
| 718 |
+
"loss": 2.2331,
|
| 719 |
+
"step": 117
|
| 720 |
+
},
|
| 721 |
+
{
|
| 722 |
+
"epoch": 0.24,
|
| 723 |
+
"learning_rate": 0.00018879554996905766,
|
| 724 |
+
"loss": 2.2188,
|
| 725 |
+
"step": 118
|
| 726 |
+
},
|
| 727 |
+
{
|
| 728 |
+
"epoch": 0.24,
|
| 729 |
+
"learning_rate": 0.00018847011335021449,
|
| 730 |
+
"loss": 2.2649,
|
| 731 |
+
"step": 119
|
| 732 |
+
},
|
| 733 |
+
{
|
| 734 |
+
"epoch": 0.24,
|
| 735 |
+
"learning_rate": 0.0001881403067509104,
|
| 736 |
+
"loss": 2.2905,
|
| 737 |
+
"step": 120
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 0.24,
|
| 741 |
+
"learning_rate": 0.00018780614646193942,
|
| 742 |
+
"loss": 2.2859,
|
| 743 |
+
"step": 121
|
| 744 |
+
},
|
| 745 |
+
{
|
| 746 |
+
"epoch": 0.25,
|
| 747 |
+
"learning_rate": 0.0001874676489891461,
|
| 748 |
+
"loss": 2.1971,
|
| 749 |
+
"step": 122
|
| 750 |
+
},
|
| 751 |
+
{
|
| 752 |
+
"epoch": 0.25,
|
| 753 |
+
"learning_rate": 0.00018712483105261005,
|
| 754 |
+
"loss": 2.2541,
|
| 755 |
+
"step": 123
|
| 756 |
+
},
|
| 757 |
+
{
|
| 758 |
+
"epoch": 0.25,
|
| 759 |
+
"learning_rate": 0.00018677770958582023,
|
| 760 |
+
"loss": 2.2179,
|
| 761 |
+
"step": 124
|
| 762 |
+
},
|
| 763 |
+
{
|
| 764 |
+
"epoch": 0.25,
|
| 765 |
+
"learning_rate": 0.00018642630173483832,
|
| 766 |
+
"loss": 2.2198,
|
| 767 |
+
"step": 125
|
| 768 |
+
},
|
| 769 |
+
{
|
| 770 |
+
"epoch": 0.25,
|
| 771 |
+
"eval_loss": 2.3039979934692383,
|
| 772 |
+
"eval_runtime": 51.9209,
|
| 773 |
+
"eval_samples_per_second": 24.21,
|
| 774 |
+
"eval_steps_per_second": 1.021,
|
| 775 |
+
"step": 125
|
| 776 |
+
},
|
| 777 |
+
{
|
| 778 |
+
"epoch": 0.25,
|
| 779 |
+
"learning_rate": 0.00018607062485745212,
|
| 780 |
+
"loss": 2.332,
|
| 781 |
+
"step": 126
|
| 782 |
+
},
|
| 783 |
+
{
|
| 784 |
+
"epoch": 0.26,
|
| 785 |
+
"learning_rate": 0.0001857106965223177,
|
| 786 |
+
"loss": 2.2744,
|
| 787 |
+
"step": 127
|
| 788 |
+
},
|
| 789 |
+
{
|
| 790 |
+
"epoch": 0.26,
|
| 791 |
+
"learning_rate": 0.00018534653450809197,
|
| 792 |
+
"loss": 2.2837,
|
| 793 |
+
"step": 128
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 0.26,
|
| 797 |
+
"learning_rate": 0.0001849781568025545,
|
| 798 |
+
"loss": 2.249,
|
| 799 |
+
"step": 129
|
| 800 |
+
},
|
| 801 |
+
{
|
| 802 |
+
"epoch": 0.26,
|
| 803 |
+
"learning_rate": 0.00018460558160171865,
|
| 804 |
+
"loss": 2.3065,
|
| 805 |
+
"step": 130
|
| 806 |
+
},
|
| 807 |
+
{
|
| 808 |
+
"epoch": 0.26,
|
| 809 |
+
"learning_rate": 0.0001842288273089332,
|
| 810 |
+
"loss": 2.2788,
|
| 811 |
+
"step": 131
|
| 812 |
+
},
|
| 813 |
+
{
|
| 814 |
+
"epoch": 0.27,
|
| 815 |
+
"learning_rate": 0.0001838479125339731,
|
| 816 |
+
"loss": 2.2946,
|
| 817 |
+
"step": 132
|
| 818 |
+
},
|
| 819 |
+
{
|
| 820 |
+
"epoch": 0.27,
|
| 821 |
+
"learning_rate": 0.00018346285609212025,
|
| 822 |
+
"loss": 2.2101,
|
| 823 |
+
"step": 133
|
| 824 |
+
},
|
| 825 |
+
{
|
| 826 |
+
"epoch": 0.27,
|
| 827 |
+
"learning_rate": 0.0001830736770032341,
|
| 828 |
+
"loss": 2.2492,
|
| 829 |
+
"step": 134
|
| 830 |
+
},
|
| 831 |
+
{
|
| 832 |
+
"epoch": 0.27,
|
| 833 |
+
"learning_rate": 0.0001826803944908124,
|
| 834 |
+
"loss": 2.1931,
|
| 835 |
+
"step": 135
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"epoch": 0.27,
|
| 839 |
+
"learning_rate": 0.00018228302798104127,
|
| 840 |
+
"loss": 2.196,
|
| 841 |
+
"step": 136
|
| 842 |
+
},
|
| 843 |
+
{
|
| 844 |
+
"epoch": 0.28,
|
| 845 |
+
"learning_rate": 0.00018188159710183594,
|
| 846 |
+
"loss": 2.2698,
|
| 847 |
+
"step": 137
|
| 848 |
+
},
|
| 849 |
+
{
|
| 850 |
+
"epoch": 0.28,
|
| 851 |
+
"learning_rate": 0.0001814761216818711,
|
| 852 |
+
"loss": 2.3152,
|
| 853 |
+
"step": 138
|
| 854 |
+
},
|
| 855 |
+
{
|
| 856 |
+
"epoch": 0.28,
|
| 857 |
+
"learning_rate": 0.00018106662174960153,
|
| 858 |
+
"loss": 2.2096,
|
| 859 |
+
"step": 139
|
| 860 |
+
},
|
| 861 |
+
{
|
| 862 |
+
"epoch": 0.28,
|
| 863 |
+
"learning_rate": 0.00018065311753227273,
|
| 864 |
+
"loss": 2.2467,
|
| 865 |
+
"step": 140
|
| 866 |
+
},
|
| 867 |
+
{
|
| 868 |
+
"epoch": 0.28,
|
| 869 |
+
"learning_rate": 0.0001802356294549218,
|
| 870 |
+
"loss": 2.2236,
|
| 871 |
+
"step": 141
|
| 872 |
+
},
|
| 873 |
+
{
|
| 874 |
+
"epoch": 0.29,
|
| 875 |
+
"learning_rate": 0.00017981417813936864,
|
| 876 |
+
"loss": 2.2537,
|
| 877 |
+
"step": 142
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"epoch": 0.29,
|
| 881 |
+
"learning_rate": 0.0001793887844031972,
|
| 882 |
+
"loss": 2.2716,
|
| 883 |
+
"step": 143
|
| 884 |
+
},
|
| 885 |
+
{
|
| 886 |
+
"epoch": 0.29,
|
| 887 |
+
"learning_rate": 0.00017895946925872733,
|
| 888 |
+
"loss": 2.2136,
|
| 889 |
+
"step": 144
|
| 890 |
+
},
|
| 891 |
+
{
|
| 892 |
+
"epoch": 0.29,
|
| 893 |
+
"learning_rate": 0.00017852625391197674,
|
| 894 |
+
"loss": 2.2374,
|
| 895 |
+
"step": 145
|
| 896 |
+
},
|
| 897 |
+
{
|
| 898 |
+
"epoch": 0.29,
|
| 899 |
+
"learning_rate": 0.00017808915976161362,
|
| 900 |
+
"loss": 2.2258,
|
| 901 |
+
"step": 146
|
| 902 |
+
},
|
| 903 |
+
{
|
| 904 |
+
"epoch": 0.3,
|
| 905 |
+
"learning_rate": 0.00017764820839789964,
|
| 906 |
+
"loss": 2.1188,
|
| 907 |
+
"step": 147
|
| 908 |
+
},
|
| 909 |
+
{
|
| 910 |
+
"epoch": 0.3,
|
| 911 |
+
"learning_rate": 0.00017720342160162348,
|
| 912 |
+
"loss": 2.1268,
|
| 913 |
+
"step": 148
|
| 914 |
+
},
|
| 915 |
+
{
|
| 916 |
+
"epoch": 0.3,
|
| 917 |
+
"learning_rate": 0.000176754821343025,
|
| 918 |
+
"loss": 2.2547,
|
| 919 |
+
"step": 149
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"epoch": 0.3,
|
| 923 |
+
"learning_rate": 0.00017630242978070997,
|
| 924 |
+
"loss": 2.314,
|
| 925 |
+
"step": 150
|
| 926 |
+
},
|
| 927 |
+
{
|
| 928 |
+
"epoch": 0.3,
|
| 929 |
+
"learning_rate": 0.00017584626926055554,
|
| 930 |
+
"loss": 2.3001,
|
| 931 |
+
"step": 151
|
| 932 |
+
},
|
| 933 |
+
{
|
| 934 |
+
"epoch": 0.31,
|
| 935 |
+
"learning_rate": 0.0001753863623146066,
|
| 936 |
+
"loss": 2.1514,
|
| 937 |
+
"step": 152
|
| 938 |
+
},
|
| 939 |
+
{
|
| 940 |
+
"epoch": 0.31,
|
| 941 |
+
"learning_rate": 0.00017492273165996261,
|
| 942 |
+
"loss": 2.2405,
|
| 943 |
+
"step": 153
|
| 944 |
+
},
|
| 945 |
+
{
|
| 946 |
+
"epoch": 0.31,
|
| 947 |
+
"learning_rate": 0.0001744554001976556,
|
| 948 |
+
"loss": 2.2064,
|
| 949 |
+
"step": 154
|
| 950 |
+
},
|
| 951 |
+
{
|
| 952 |
+
"epoch": 0.31,
|
| 953 |
+
"learning_rate": 0.00017398439101151905,
|
| 954 |
+
"loss": 2.1977,
|
| 955 |
+
"step": 155
|
| 956 |
+
},
|
| 957 |
+
{
|
| 958 |
+
"epoch": 0.31,
|
| 959 |
+
"learning_rate": 0.0001735097273670475,
|
| 960 |
+
"loss": 2.2881,
|
| 961 |
+
"step": 156
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"epoch": 0.32,
|
| 965 |
+
"learning_rate": 0.00017303143271024744,
|
| 966 |
+
"loss": 2.1925,
|
| 967 |
+
"step": 157
|
| 968 |
+
},
|
| 969 |
+
{
|
| 970 |
+
"epoch": 0.32,
|
| 971 |
+
"learning_rate": 0.00017254953066647913,
|
| 972 |
+
"loss": 2.1888,
|
| 973 |
+
"step": 158
|
| 974 |
+
},
|
| 975 |
+
{
|
| 976 |
+
"epoch": 0.32,
|
| 977 |
+
"learning_rate": 0.0001720640450392898,
|
| 978 |
+
"loss": 2.2044,
|
| 979 |
+
"step": 159
|
| 980 |
+
},
|
| 981 |
+
{
|
| 982 |
+
"epoch": 0.32,
|
| 983 |
+
"learning_rate": 0.00017157499980923767,
|
| 984 |
+
"loss": 2.145,
|
| 985 |
+
"step": 160
|
| 986 |
+
},
|
| 987 |
+
{
|
| 988 |
+
"epoch": 0.32,
|
| 989 |
+
"learning_rate": 0.0001710824191327075,
|
| 990 |
+
"loss": 2.1138,
|
| 991 |
+
"step": 161
|
| 992 |
+
},
|
| 993 |
+
{
|
| 994 |
+
"epoch": 0.33,
|
| 995 |
+
"learning_rate": 0.0001705863273407174,
|
| 996 |
+
"loss": 2.2124,
|
| 997 |
+
"step": 162
|
| 998 |
+
},
|
| 999 |
+
{
|
| 1000 |
+
"epoch": 0.33,
|
| 1001 |
+
"learning_rate": 0.00017008674893771706,
|
| 1002 |
+
"loss": 2.0964,
|
| 1003 |
+
"step": 163
|
| 1004 |
+
},
|
| 1005 |
+
{
|
| 1006 |
+
"epoch": 0.33,
|
| 1007 |
+
"learning_rate": 0.00016958370860037717,
|
| 1008 |
+
"loss": 2.208,
|
| 1009 |
+
"step": 164
|
| 1010 |
+
},
|
| 1011 |
+
{
|
| 1012 |
+
"epoch": 0.33,
|
| 1013 |
+
"learning_rate": 0.00016907723117637083,
|
| 1014 |
+
"loss": 2.1633,
|
| 1015 |
+
"step": 165
|
| 1016 |
+
},
|
| 1017 |
+
{
|
| 1018 |
+
"epoch": 0.33,
|
| 1019 |
+
"learning_rate": 0.00016856734168314583,
|
| 1020 |
+
"loss": 2.1419,
|
| 1021 |
+
"step": 166
|
| 1022 |
+
},
|
| 1023 |
+
{
|
| 1024 |
+
"epoch": 0.34,
|
| 1025 |
+
"learning_rate": 0.0001680540653066891,
|
| 1026 |
+
"loss": 2.1253,
|
| 1027 |
+
"step": 167
|
| 1028 |
+
},
|
| 1029 |
+
{
|
| 1030 |
+
"epoch": 0.34,
|
| 1031 |
+
"learning_rate": 0.00016753742740028277,
|
| 1032 |
+
"loss": 2.1568,
|
| 1033 |
+
"step": 168
|
| 1034 |
+
},
|
| 1035 |
+
{
|
| 1036 |
+
"epoch": 0.34,
|
| 1037 |
+
"learning_rate": 0.00016701745348325156,
|
| 1038 |
+
"loss": 2.1131,
|
| 1039 |
+
"step": 169
|
| 1040 |
+
},
|
| 1041 |
+
{
|
| 1042 |
+
"epoch": 0.34,
|
| 1043 |
+
"learning_rate": 0.0001664941692397025,
|
| 1044 |
+
"loss": 2.1684,
|
| 1045 |
+
"step": 170
|
| 1046 |
+
},
|
| 1047 |
+
{
|
| 1048 |
+
"epoch": 0.34,
|
| 1049 |
+
"learning_rate": 0.00016596760051725609,
|
| 1050 |
+
"loss": 2.1378,
|
| 1051 |
+
"step": 171
|
| 1052 |
+
},
|
| 1053 |
+
{
|
| 1054 |
+
"epoch": 0.35,
|
| 1055 |
+
"learning_rate": 0.00016543777332576976,
|
| 1056 |
+
"loss": 2.0971,
|
| 1057 |
+
"step": 172
|
| 1058 |
+
},
|
| 1059 |
+
{
|
| 1060 |
+
"epoch": 0.35,
|
| 1061 |
+
"learning_rate": 0.00016490471383605288,
|
| 1062 |
+
"loss": 2.1363,
|
| 1063 |
+
"step": 173
|
| 1064 |
+
},
|
| 1065 |
+
{
|
| 1066 |
+
"epoch": 0.35,
|
| 1067 |
+
"learning_rate": 0.00016436844837857417,
|
| 1068 |
+
"loss": 2.0525,
|
| 1069 |
+
"step": 174
|
| 1070 |
+
},
|
| 1071 |
+
{
|
| 1072 |
+
"epoch": 0.35,
|
| 1073 |
+
"learning_rate": 0.00016382900344216115,
|
| 1074 |
+
"loss": 2.1323,
|
| 1075 |
+
"step": 175
|
| 1076 |
+
},
|
| 1077 |
+
{
|
| 1078 |
+
"epoch": 0.35,
|
| 1079 |
+
"learning_rate": 0.0001632864056726917,
|
| 1080 |
+
"loss": 2.1257,
|
| 1081 |
+
"step": 176
|
| 1082 |
+
},
|
| 1083 |
+
{
|
| 1084 |
+
"epoch": 0.36,
|
| 1085 |
+
"learning_rate": 0.00016274068187177771,
|
| 1086 |
+
"loss": 2.0854,
|
| 1087 |
+
"step": 177
|
| 1088 |
+
},
|
| 1089 |
+
{
|
| 1090 |
+
"epoch": 0.36,
|
| 1091 |
+
"learning_rate": 0.00016219185899544154,
|
| 1092 |
+
"loss": 2.2714,
|
| 1093 |
+
"step": 178
|
| 1094 |
+
},
|
| 1095 |
+
{
|
| 1096 |
+
"epoch": 0.36,
|
| 1097 |
+
"learning_rate": 0.00016163996415278424,
|
| 1098 |
+
"loss": 2.0961,
|
| 1099 |
+
"step": 179
|
| 1100 |
+
},
|
| 1101 |
+
{
|
| 1102 |
+
"epoch": 0.36,
|
| 1103 |
+
"learning_rate": 0.00016108502460464666,
|
| 1104 |
+
"loss": 2.1213,
|
| 1105 |
+
"step": 180
|
| 1106 |
+
},
|
| 1107 |
+
{
|
| 1108 |
+
"epoch": 0.36,
|
| 1109 |
+
"learning_rate": 0.00016052706776226286,
|
| 1110 |
+
"loss": 2.121,
|
| 1111 |
+
"step": 181
|
| 1112 |
+
},
|
| 1113 |
+
{
|
| 1114 |
+
"epoch": 0.37,
|
| 1115 |
+
"learning_rate": 0.00015996612118590603,
|
| 1116 |
+
"loss": 2.1398,
|
| 1117 |
+
"step": 182
|
| 1118 |
+
},
|
| 1119 |
+
{
|
| 1120 |
+
"epoch": 0.37,
|
| 1121 |
+
"learning_rate": 0.00015940221258352742,
|
| 1122 |
+
"loss": 2.0577,
|
| 1123 |
+
"step": 183
|
| 1124 |
+
},
|
| 1125 |
+
{
|
| 1126 |
+
"epoch": 0.37,
|
| 1127 |
+
"learning_rate": 0.00015883536980938734,
|
| 1128 |
+
"loss": 2.0841,
|
| 1129 |
+
"step": 184
|
| 1130 |
+
},
|
| 1131 |
+
{
|
| 1132 |
+
"epoch": 0.37,
|
| 1133 |
+
"learning_rate": 0.00015826562086267956,
|
| 1134 |
+
"loss": 2.0541,
|
| 1135 |
+
"step": 185
|
| 1136 |
+
},
|
| 1137 |
+
{
|
| 1138 |
+
"epoch": 0.37,
|
| 1139 |
+
"learning_rate": 0.00015769299388614826,
|
| 1140 |
+
"loss": 2.0365,
|
| 1141 |
+
"step": 186
|
| 1142 |
+
},
|
| 1143 |
+
{
|
| 1144 |
+
"epoch": 0.38,
|
| 1145 |
+
"learning_rate": 0.00015711751716469786,
|
| 1146 |
+
"loss": 2.1263,
|
| 1147 |
+
"step": 187
|
| 1148 |
+
},
|
| 1149 |
+
{
|
| 1150 |
+
"epoch": 0.38,
|
| 1151 |
+
"learning_rate": 0.00015653921912399589,
|
| 1152 |
+
"loss": 2.0934,
|
| 1153 |
+
"step": 188
|
| 1154 |
+
},
|
| 1155 |
+
{
|
| 1156 |
+
"epoch": 0.38,
|
| 1157 |
+
"learning_rate": 0.0001559581283290689,
|
| 1158 |
+
"loss": 2.0882,
|
| 1159 |
+
"step": 189
|
| 1160 |
+
},
|
| 1161 |
+
{
|
| 1162 |
+
"epoch": 0.38,
|
| 1163 |
+
"learning_rate": 0.00015537427348289153,
|
| 1164 |
+
"loss": 1.9549,
|
| 1165 |
+
"step": 190
|
| 1166 |
+
},
|
| 1167 |
+
{
|
| 1168 |
+
"epoch": 0.38,
|
| 1169 |
+
"learning_rate": 0.0001547876834249687,
|
| 1170 |
+
"loss": 2.1164,
|
| 1171 |
+
"step": 191
|
| 1172 |
+
},
|
| 1173 |
+
{
|
| 1174 |
+
"epoch": 0.39,
|
| 1175 |
+
"learning_rate": 0.0001541983871299111,
|
| 1176 |
+
"loss": 2.0739,
|
| 1177 |
+
"step": 192
|
| 1178 |
+
},
|
| 1179 |
+
{
|
| 1180 |
+
"epoch": 0.39,
|
| 1181 |
+
"learning_rate": 0.000153606413706004,
|
| 1182 |
+
"loss": 2.0541,
|
| 1183 |
+
"step": 193
|
| 1184 |
+
},
|
| 1185 |
+
{
|
| 1186 |
+
"epoch": 0.39,
|
| 1187 |
+
"learning_rate": 0.00015301179239376938,
|
| 1188 |
+
"loss": 1.9866,
|
| 1189 |
+
"step": 194
|
| 1190 |
+
},
|
| 1191 |
+
{
|
| 1192 |
+
"epoch": 0.39,
|
| 1193 |
+
"learning_rate": 0.0001524145525645216,
|
| 1194 |
+
"loss": 2.0699,
|
| 1195 |
+
"step": 195
|
| 1196 |
+
},
|
| 1197 |
+
{
|
| 1198 |
+
"epoch": 0.39,
|
| 1199 |
+
"learning_rate": 0.00015181472371891686,
|
| 1200 |
+
"loss": 2.0908,
|
| 1201 |
+
"step": 196
|
| 1202 |
+
},
|
| 1203 |
+
{
|
| 1204 |
+
"epoch": 0.4,
|
| 1205 |
+
"learning_rate": 0.0001512123354854955,
|
| 1206 |
+
"loss": 1.968,
|
| 1207 |
+
"step": 197
|
| 1208 |
+
},
|
| 1209 |
+
{
|
| 1210 |
+
"epoch": 0.4,
|
| 1211 |
+
"learning_rate": 0.00015060741761921902,
|
| 1212 |
+
"loss": 2.0859,
|
| 1213 |
+
"step": 198
|
| 1214 |
+
},
|
| 1215 |
+
{
|
| 1216 |
+
"epoch": 0.4,
|
| 1217 |
+
"learning_rate": 0.00015000000000000001,
|
| 1218 |
+
"loss": 1.9517,
|
| 1219 |
+
"step": 199
|
| 1220 |
+
},
|
| 1221 |
+
{
|
| 1222 |
+
"epoch": 0.4,
|
| 1223 |
+
"learning_rate": 0.00014939011263122634,
|
| 1224 |
+
"loss": 2.0148,
|
| 1225 |
+
"step": 200
|
| 1226 |
+
},
|
| 1227 |
+
{
|
| 1228 |
+
"epoch": 0.4,
|
| 1229 |
+
"learning_rate": 0.00014877778563827923,
|
| 1230 |
+
"loss": 2.0794,
|
| 1231 |
+
"step": 201
|
| 1232 |
+
},
|
| 1233 |
+
{
|
| 1234 |
+
"epoch": 0.41,
|
| 1235 |
+
"learning_rate": 0.000148163049267045,
|
| 1236 |
+
"loss": 1.9808,
|
| 1237 |
+
"step": 202
|
| 1238 |
+
},
|
| 1239 |
+
{
|
| 1240 |
+
"epoch": 0.41,
|
| 1241 |
+
"learning_rate": 0.00014754593388242117,
|
| 1242 |
+
"loss": 2.0028,
|
| 1243 |
+
"step": 203
|
| 1244 |
+
},
|
| 1245 |
+
{
|
| 1246 |
+
"epoch": 0.41,
|
| 1247 |
+
"learning_rate": 0.0001469264699668168,
|
| 1248 |
+
"loss": 2.14,
|
| 1249 |
+
"step": 204
|
| 1250 |
+
},
|
| 1251 |
+
{
|
| 1252 |
+
"epoch": 0.41,
|
| 1253 |
+
"learning_rate": 0.00014630468811864633,
|
| 1254 |
+
"loss": 2.0264,
|
| 1255 |
+
"step": 205
|
| 1256 |
+
},
|
| 1257 |
+
{
|
| 1258 |
+
"epoch": 0.41,
|
| 1259 |
+
"learning_rate": 0.00014568061905081875,
|
| 1260 |
+
"loss": 1.9986,
|
| 1261 |
+
"step": 206
|
| 1262 |
+
},
|
| 1263 |
+
{
|
| 1264 |
+
"epoch": 0.42,
|
| 1265 |
+
"learning_rate": 0.00014505429358922,
|
| 1266 |
+
"loss": 2.0302,
|
| 1267 |
+
"step": 207
|
| 1268 |
+
},
|
| 1269 |
+
{
|
| 1270 |
+
"epoch": 0.42,
|
| 1271 |
+
"learning_rate": 0.00014442574267119074,
|
| 1272 |
+
"loss": 1.894,
|
| 1273 |
+
"step": 208
|
| 1274 |
+
},
|
| 1275 |
+
{
|
| 1276 |
+
"epoch": 0.42,
|
| 1277 |
+
"learning_rate": 0.00014379499734399798,
|
| 1278 |
+
"loss": 2.0404,
|
| 1279 |
+
"step": 209
|
| 1280 |
+
},
|
| 1281 |
+
{
|
| 1282 |
+
"epoch": 0.42,
|
| 1283 |
+
"learning_rate": 0.00014316208876330146,
|
| 1284 |
+
"loss": 1.8718,
|
| 1285 |
+
"step": 210
|
| 1286 |
+
},
|
| 1287 |
+
{
|
| 1288 |
+
"epoch": 0.42,
|
| 1289 |
+
"learning_rate": 0.00014252704819161496,
|
| 1290 |
+
"loss": 1.9787,
|
| 1291 |
+
"step": 211
|
| 1292 |
+
},
|
| 1293 |
+
{
|
| 1294 |
+
"epoch": 0.43,
|
| 1295 |
+
"learning_rate": 0.00014188990699676184,
|
| 1296 |
+
"loss": 2.0504,
|
| 1297 |
+
"step": 212
|
| 1298 |
+
},
|
| 1299 |
+
{
|
| 1300 |
+
"epoch": 0.43,
|
| 1301 |
+
"learning_rate": 0.00014125069665032574,
|
| 1302 |
+
"loss": 1.9288,
|
| 1303 |
+
"step": 213
|
| 1304 |
+
},
|
| 1305 |
+
{
|
| 1306 |
+
"epoch": 0.43,
|
| 1307 |
+
"learning_rate": 0.00014060944872609606,
|
| 1308 |
+
"loss": 2.0359,
|
| 1309 |
+
"step": 214
|
| 1310 |
+
},
|
| 1311 |
+
{
|
| 1312 |
+
"epoch": 0.43,
|
| 1313 |
+
"learning_rate": 0.00013996619489850822,
|
| 1314 |
+
"loss": 2.0282,
|
| 1315 |
+
"step": 215
|
| 1316 |
+
},
|
| 1317 |
+
{
|
| 1318 |
+
"epoch": 0.43,
|
| 1319 |
+
"learning_rate": 0.0001393209669410794,
|
| 1320 |
+
"loss": 1.9534,
|
| 1321 |
+
"step": 216
|
| 1322 |
+
},
|
| 1323 |
+
{
|
| 1324 |
+
"epoch": 0.44,
|
| 1325 |
+
"learning_rate": 0.0001386737967248388,
|
| 1326 |
+
"loss": 1.9479,
|
| 1327 |
+
"step": 217
|
| 1328 |
+
},
|
| 1329 |
+
{
|
| 1330 |
+
"epoch": 0.44,
|
| 1331 |
+
"learning_rate": 0.00013802471621675338,
|
| 1332 |
+
"loss": 1.9512,
|
| 1333 |
+
"step": 218
|
| 1334 |
+
},
|
| 1335 |
+
{
|
| 1336 |
+
"epoch": 0.44,
|
| 1337 |
+
"learning_rate": 0.00013737375747814915,
|
| 1338 |
+
"loss": 1.9571,
|
| 1339 |
+
"step": 219
|
| 1340 |
+
},
|
| 1341 |
+
{
|
| 1342 |
+
"epoch": 0.44,
|
| 1343 |
+
"learning_rate": 0.0001367209526631272,
|
| 1344 |
+
"loss": 1.9223,
|
| 1345 |
+
"step": 220
|
| 1346 |
+
},
|
| 1347 |
+
{
|
| 1348 |
+
"epoch": 0.44,
|
| 1349 |
+
"learning_rate": 0.00013606633401697557,
|
| 1350 |
+
"loss": 1.9252,
|
| 1351 |
+
"step": 221
|
| 1352 |
+
},
|
| 1353 |
+
{
|
| 1354 |
+
"epoch": 0.45,
|
| 1355 |
+
"learning_rate": 0.0001354099338745764,
|
| 1356 |
+
"loss": 1.9298,
|
| 1357 |
+
"step": 222
|
| 1358 |
+
},
|
| 1359 |
+
{
|
| 1360 |
+
"epoch": 0.45,
|
| 1361 |
+
"learning_rate": 0.0001347517846588089,
|
| 1362 |
+
"loss": 1.9366,
|
| 1363 |
+
"step": 223
|
| 1364 |
+
},
|
| 1365 |
+
{
|
| 1366 |
+
"epoch": 0.45,
|
| 1367 |
+
"learning_rate": 0.0001340919188789477,
|
| 1368 |
+
"loss": 1.8472,
|
| 1369 |
+
"step": 224
|
| 1370 |
+
},
|
| 1371 |
+
{
|
| 1372 |
+
"epoch": 0.45,
|
| 1373 |
+
"learning_rate": 0.00013343036912905718,
|
| 1374 |
+
"loss": 1.9318,
|
| 1375 |
+
"step": 225
|
| 1376 |
+
},
|
| 1377 |
+
{
|
| 1378 |
+
"epoch": 0.45,
|
| 1379 |
+
"learning_rate": 0.00013276716808638126,
|
| 1380 |
+
"loss": 2.0005,
|
| 1381 |
+
"step": 226
|
| 1382 |
+
},
|
| 1383 |
+
{
|
| 1384 |
+
"epoch": 0.46,
|
| 1385 |
+
"learning_rate": 0.00013210234850972964,
|
| 1386 |
+
"loss": 1.9785,
|
| 1387 |
+
"step": 227
|
| 1388 |
+
},
|
| 1389 |
+
{
|
| 1390 |
+
"epoch": 0.46,
|
| 1391 |
+
"learning_rate": 0.00013143594323785927,
|
| 1392 |
+
"loss": 1.8198,
|
| 1393 |
+
"step": 228
|
| 1394 |
+
},
|
| 1395 |
+
{
|
| 1396 |
+
"epoch": 0.46,
|
| 1397 |
+
"learning_rate": 0.00013076798518785274,
|
| 1398 |
+
"loss": 1.8197,
|
| 1399 |
+
"step": 229
|
| 1400 |
+
},
|
| 1401 |
+
{
|
| 1402 |
+
"epoch": 0.46,
|
| 1403 |
+
"learning_rate": 0.0001300985073534919,
|
| 1404 |
+
"loss": 1.8909,
|
| 1405 |
+
"step": 230
|
| 1406 |
+
},
|
| 1407 |
+
{
|
| 1408 |
+
"epoch": 0.46,
|
| 1409 |
+
"learning_rate": 0.0001294275428036284,
|
| 1410 |
+
"loss": 1.937,
|
| 1411 |
+
"step": 231
|
| 1412 |
+
},
|
| 1413 |
+
{
|
| 1414 |
+
"epoch": 0.47,
|
| 1415 |
+
"learning_rate": 0.00012875512468055024,
|
| 1416 |
+
"loss": 1.8939,
|
| 1417 |
+
"step": 232
|
| 1418 |
+
},
|
| 1419 |
+
{
|
| 1420 |
+
"epoch": 0.47,
|
| 1421 |
+
"learning_rate": 0.00012808128619834461,
|
| 1422 |
+
"loss": 1.8948,
|
| 1423 |
+
"step": 233
|
| 1424 |
+
},
|
| 1425 |
+
{
|
| 1426 |
+
"epoch": 0.47,
|
| 1427 |
+
"learning_rate": 0.00012740606064125736,
|
| 1428 |
+
"loss": 1.9238,
|
| 1429 |
+
"step": 234
|
| 1430 |
+
},
|
| 1431 |
+
{
|
| 1432 |
+
"epoch": 0.47,
|
| 1433 |
+
"learning_rate": 0.00012672948136204887,
|
| 1434 |
+
"loss": 1.891,
|
| 1435 |
+
"step": 235
|
| 1436 |
+
},
|
| 1437 |
+
{
|
| 1438 |
+
"epoch": 0.47,
|
| 1439 |
+
"learning_rate": 0.00012605158178034654,
|
| 1440 |
+
"loss": 1.8728,
|
| 1441 |
+
"step": 236
|
| 1442 |
+
},
|
| 1443 |
+
{
|
| 1444 |
+
"epoch": 0.48,
|
| 1445 |
+
"learning_rate": 0.00012537239538099425,
|
| 1446 |
+
"loss": 1.9132,
|
| 1447 |
+
"step": 237
|
| 1448 |
+
},
|
| 1449 |
+
{
|
| 1450 |
+
"epoch": 0.48,
|
| 1451 |
+
"learning_rate": 0.0001246919557123981,
|
| 1452 |
+
"loss": 1.8633,
|
| 1453 |
+
"step": 238
|
| 1454 |
+
},
|
| 1455 |
+
{
|
| 1456 |
+
"epoch": 0.48,
|
| 1457 |
+
"learning_rate": 0.00012401029638486953,
|
| 1458 |
+
"loss": 1.8216,
|
| 1459 |
+
"step": 239
|
| 1460 |
+
},
|
| 1461 |
+
{
|
| 1462 |
+
"epoch": 0.48,
|
| 1463 |
+
"learning_rate": 0.00012332745106896482,
|
| 1464 |
+
"loss": 1.8434,
|
| 1465 |
+
"step": 240
|
| 1466 |
+
},
|
| 1467 |
+
{
|
| 1468 |
+
"epoch": 0.48,
|
| 1469 |
+
"learning_rate": 0.00012264345349382238,
|
| 1470 |
+
"loss": 1.8116,
|
| 1471 |
+
"step": 241
|
| 1472 |
+
},
|
| 1473 |
+
{
|
| 1474 |
+
"epoch": 0.49,
|
| 1475 |
+
"learning_rate": 0.0001219583374454963,
|
| 1476 |
+
"loss": 1.8152,
|
| 1477 |
+
"step": 242
|
| 1478 |
+
},
|
| 1479 |
+
{
|
| 1480 |
+
"epoch": 0.49,
|
| 1481 |
+
"learning_rate": 0.00012127213676528768,
|
| 1482 |
+
"loss": 1.8103,
|
| 1483 |
+
"step": 243
|
| 1484 |
+
},
|
| 1485 |
+
{
|
| 1486 |
+
"epoch": 0.49,
|
| 1487 |
+
"learning_rate": 0.00012058488534807303,
|
| 1488 |
+
"loss": 1.8576,
|
| 1489 |
+
"step": 244
|
| 1490 |
+
},
|
| 1491 |
+
{
|
| 1492 |
+
"epoch": 0.49,
|
| 1493 |
+
"learning_rate": 0.00011989661714062999,
|
| 1494 |
+
"loss": 1.8415,
|
| 1495 |
+
"step": 245
|
| 1496 |
+
},
|
| 1497 |
+
{
|
| 1498 |
+
"epoch": 0.49,
|
| 1499 |
+
"learning_rate": 0.00011920736613996046,
|
| 1500 |
+
"loss": 1.8708,
|
| 1501 |
+
"step": 246
|
| 1502 |
+
},
|
| 1503 |
+
{
|
| 1504 |
+
"epoch": 0.5,
|
| 1505 |
+
"learning_rate": 0.00011851716639161159,
|
| 1506 |
+
"loss": 1.8049,
|
| 1507 |
+
"step": 247
|
| 1508 |
+
},
|
| 1509 |
+
{
|
| 1510 |
+
"epoch": 0.5,
|
| 1511 |
+
"learning_rate": 0.0001178260519879937,
|
| 1512 |
+
"loss": 1.8555,
|
| 1513 |
+
"step": 248
|
| 1514 |
+
},
|
| 1515 |
+
{
|
| 1516 |
+
"epoch": 0.5,
|
| 1517 |
+
"learning_rate": 0.00011713405706669667,
|
| 1518 |
+
"loss": 1.8647,
|
| 1519 |
+
"step": 249
|
| 1520 |
+
},
|
| 1521 |
+
{
|
| 1522 |
+
"epoch": 0.5,
|
| 1523 |
+
"learning_rate": 0.00011644121580880345,
|
| 1524 |
+
"loss": 1.8488,
|
| 1525 |
+
"step": 250
|
| 1526 |
+
},
|
| 1527 |
+
{
|
| 1528 |
+
"epoch": 0.5,
|
| 1529 |
+
"eval_loss": 1.8219653367996216,
|
| 1530 |
+
"eval_runtime": 51.9603,
|
| 1531 |
+
"eval_samples_per_second": 24.192,
|
| 1532 |
+
"eval_steps_per_second": 1.02,
|
| 1533 |
+
"step": 250
|
| 1534 |
+
},
|
| 1535 |
+
{
|
| 1536 |
+
"epoch": 0.5,
|
| 1537 |
+
"learning_rate": 0.0001157475624372018,
|
| 1538 |
+
"loss": 1.8457,
|
| 1539 |
+
"step": 251
|
| 1540 |
+
},
|
| 1541 |
+
{
|
| 1542 |
+
"epoch": 0.51,
|
| 1543 |
+
"learning_rate": 0.00011505313121489383,
|
| 1544 |
+
"loss": 1.7723,
|
| 1545 |
+
"step": 252
|
| 1546 |
+
},
|
| 1547 |
+
{
|
| 1548 |
+
"epoch": 0.51,
|
| 1549 |
+
"learning_rate": 0.00011435795644330359,
|
| 1550 |
+
"loss": 1.7381,
|
| 1551 |
+
"step": 253
|
| 1552 |
+
},
|
| 1553 |
+
{
|
| 1554 |
+
"epoch": 0.51,
|
| 1555 |
+
"learning_rate": 0.0001136620724605827,
|
| 1556 |
+
"loss": 1.9117,
|
| 1557 |
+
"step": 254
|
| 1558 |
+
},
|
| 1559 |
+
{
|
| 1560 |
+
"epoch": 0.51,
|
| 1561 |
+
"learning_rate": 0.00011296551363991432,
|
| 1562 |
+
"loss": 1.8667,
|
| 1563 |
+
"step": 255
|
| 1564 |
+
},
|
| 1565 |
+
{
|
| 1566 |
+
"epoch": 0.51,
|
| 1567 |
+
"learning_rate": 0.00011226831438781518,
|
| 1568 |
+
"loss": 1.7878,
|
| 1569 |
+
"step": 256
|
| 1570 |
+
},
|
| 1571 |
+
{
|
| 1572 |
+
"epoch": 0.52,
|
| 1573 |
+
"learning_rate": 0.00011157050914243614,
|
| 1574 |
+
"loss": 1.7116,
|
| 1575 |
+
"step": 257
|
| 1576 |
+
},
|
| 1577 |
+
{
|
| 1578 |
+
"epoch": 0.52,
|
| 1579 |
+
"learning_rate": 0.00011087213237186108,
|
| 1580 |
+
"loss": 1.8073,
|
| 1581 |
+
"step": 258
|
| 1582 |
+
},
|
| 1583 |
+
{
|
| 1584 |
+
"epoch": 0.52,
|
| 1585 |
+
"learning_rate": 0.00011017321857240432,
|
| 1586 |
+
"loss": 1.7985,
|
| 1587 |
+
"step": 259
|
| 1588 |
+
},
|
| 1589 |
+
{
|
| 1590 |
+
"epoch": 0.52,
|
| 1591 |
+
"learning_rate": 0.00010947380226690684,
|
| 1592 |
+
"loss": 1.8121,
|
| 1593 |
+
"step": 260
|
| 1594 |
+
},
|
| 1595 |
+
{
|
| 1596 |
+
"epoch": 0.52,
|
| 1597 |
+
"learning_rate": 0.00010877391800303074,
|
| 1598 |
+
"loss": 1.6811,
|
| 1599 |
+
"step": 261
|
| 1600 |
+
},
|
| 1601 |
+
{
|
| 1602 |
+
"epoch": 0.53,
|
| 1603 |
+
"learning_rate": 0.00010807360035155305,
|
| 1604 |
+
"loss": 1.7594,
|
| 1605 |
+
"step": 262
|
| 1606 |
+
},
|
| 1607 |
+
{
|
| 1608 |
+
"epoch": 0.53,
|
| 1609 |
+
"learning_rate": 0.00010737288390465792,
|
| 1610 |
+
"loss": 1.7029,
|
| 1611 |
+
"step": 263
|
| 1612 |
+
},
|
| 1613 |
+
{
|
| 1614 |
+
"epoch": 0.53,
|
| 1615 |
+
"learning_rate": 0.00010667180327422797,
|
| 1616 |
+
"loss": 1.7555,
|
| 1617 |
+
"step": 264
|
| 1618 |
+
},
|
| 1619 |
+
{
|
| 1620 |
+
"epoch": 0.53,
|
| 1621 |
+
"learning_rate": 0.00010597039309013472,
|
| 1622 |
+
"loss": 1.8195,
|
| 1623 |
+
"step": 265
|
| 1624 |
+
},
|
| 1625 |
+
{
|
| 1626 |
+
"epoch": 0.53,
|
| 1627 |
+
"learning_rate": 0.00010526868799852796,
|
| 1628 |
+
"loss": 1.7216,
|
| 1629 |
+
"step": 266
|
| 1630 |
+
},
|
| 1631 |
+
{
|
| 1632 |
+
"epoch": 0.54,
|
| 1633 |
+
"learning_rate": 0.00010456672266012446,
|
| 1634 |
+
"loss": 1.7184,
|
| 1635 |
+
"step": 267
|
| 1636 |
+
},
|
| 1637 |
+
{
|
| 1638 |
+
"epoch": 0.54,
|
| 1639 |
+
"learning_rate": 0.00010386453174849584,
|
| 1640 |
+
"loss": 1.7327,
|
| 1641 |
+
"step": 268
|
| 1642 |
+
},
|
| 1643 |
+
{
|
| 1644 |
+
"epoch": 0.54,
|
| 1645 |
+
"learning_rate": 0.0001031621499483559,
|
| 1646 |
+
"loss": 1.7251,
|
| 1647 |
+
"step": 269
|
| 1648 |
+
},
|
| 1649 |
+
{
|
| 1650 |
+
"epoch": 0.54,
|
| 1651 |
+
"learning_rate": 0.00010245961195384743,
|
| 1652 |
+
"loss": 1.7582,
|
| 1653 |
+
"step": 270
|
| 1654 |
+
},
|
| 1655 |
+
{
|
| 1656 |
+
"epoch": 0.54,
|
| 1657 |
+
"learning_rate": 0.00010175695246682841,
|
| 1658 |
+
"loss": 1.7147,
|
| 1659 |
+
"step": 271
|
| 1660 |
+
},
|
| 1661 |
+
{
|
| 1662 |
+
"epoch": 0.55,
|
| 1663 |
+
"learning_rate": 0.00010105420619515798,
|
| 1664 |
+
"loss": 1.7322,
|
| 1665 |
+
"step": 272
|
| 1666 |
+
},
|
| 1667 |
+
{
|
| 1668 |
+
"epoch": 0.55,
|
| 1669 |
+
"learning_rate": 0.00010035140785098187,
|
| 1670 |
+
"loss": 1.7607,
|
| 1671 |
+
"step": 273
|
| 1672 |
+
},
|
| 1673 |
+
{
|
| 1674 |
+
"epoch": 0.55,
|
| 1675 |
+
"learning_rate": 9.964859214901814e-05,
|
| 1676 |
+
"loss": 1.7381,
|
| 1677 |
+
"step": 274
|
| 1678 |
+
},
|
| 1679 |
+
{
|
| 1680 |
+
"epoch": 0.55,
|
| 1681 |
+
"learning_rate": 9.894579380484204e-05,
|
| 1682 |
+
"loss": 1.7026,
|
| 1683 |
+
"step": 275
|
| 1684 |
+
},
|
| 1685 |
+
{
|
| 1686 |
+
"epoch": 0.55,
|
| 1687 |
+
"learning_rate": 9.82430475331716e-05,
|
| 1688 |
+
"loss": 1.6587,
|
| 1689 |
+
"step": 276
|
| 1690 |
+
},
|
| 1691 |
+
{
|
| 1692 |
+
"epoch": 0.56,
|
| 1693 |
+
"learning_rate": 9.754038804615257e-05,
|
| 1694 |
+
"loss": 1.7142,
|
| 1695 |
+
"step": 277
|
| 1696 |
+
},
|
| 1697 |
+
{
|
| 1698 |
+
"epoch": 0.56,
|
| 1699 |
+
"learning_rate": 9.683785005164411e-05,
|
| 1700 |
+
"loss": 1.738,
|
| 1701 |
+
"step": 278
|
| 1702 |
+
},
|
| 1703 |
+
{
|
| 1704 |
+
"epoch": 0.56,
|
| 1705 |
+
"learning_rate": 9.613546825150421e-05,
|
| 1706 |
+
"loss": 1.672,
|
| 1707 |
+
"step": 279
|
| 1708 |
+
},
|
| 1709 |
+
{
|
| 1710 |
+
"epoch": 0.56,
|
| 1711 |
+
"learning_rate": 9.543327733987557e-05,
|
| 1712 |
+
"loss": 1.7901,
|
| 1713 |
+
"step": 280
|
| 1714 |
+
},
|
| 1715 |
+
{
|
| 1716 |
+
"epoch": 0.56,
|
| 1717 |
+
"learning_rate": 9.473131200147205e-05,
|
| 1718 |
+
"loss": 1.7879,
|
| 1719 |
+
"step": 281
|
| 1720 |
+
},
|
| 1721 |
+
{
|
| 1722 |
+
"epoch": 0.57,
|
| 1723 |
+
"learning_rate": 9.402960690986532e-05,
|
| 1724 |
+
"loss": 1.6247,
|
| 1725 |
+
"step": 282
|
| 1726 |
+
},
|
| 1727 |
+
{
|
| 1728 |
+
"epoch": 0.57,
|
| 1729 |
+
"learning_rate": 9.332819672577206e-05,
|
| 1730 |
+
"loss": 1.7031,
|
| 1731 |
+
"step": 283
|
| 1732 |
+
},
|
| 1733 |
+
{
|
| 1734 |
+
"epoch": 0.57,
|
| 1735 |
+
"learning_rate": 9.26271160953421e-05,
|
| 1736 |
+
"loss": 1.72,
|
| 1737 |
+
"step": 284
|
| 1738 |
+
},
|
| 1739 |
+
{
|
| 1740 |
+
"epoch": 0.57,
|
| 1741 |
+
"learning_rate": 9.192639964844695e-05,
|
| 1742 |
+
"loss": 1.7805,
|
| 1743 |
+
"step": 285
|
| 1744 |
+
},
|
| 1745 |
+
{
|
| 1746 |
+
"epoch": 0.57,
|
| 1747 |
+
"learning_rate": 9.122608199696928e-05,
|
| 1748 |
+
"loss": 1.7854,
|
| 1749 |
+
"step": 286
|
| 1750 |
+
},
|
| 1751 |
+
{
|
| 1752 |
+
"epoch": 0.58,
|
| 1753 |
+
"learning_rate": 9.052619773309317e-05,
|
| 1754 |
+
"loss": 1.6424,
|
| 1755 |
+
"step": 287
|
| 1756 |
+
},
|
| 1757 |
+
{
|
| 1758 |
+
"epoch": 0.58,
|
| 1759 |
+
"learning_rate": 8.982678142759566e-05,
|
| 1760 |
+
"loss": 1.6995,
|
| 1761 |
+
"step": 288
|
| 1762 |
+
},
|
| 1763 |
+
{
|
| 1764 |
+
"epoch": 0.58,
|
| 1765 |
+
"learning_rate": 8.912786762813893e-05,
|
| 1766 |
+
"loss": 1.6633,
|
| 1767 |
+
"step": 289
|
| 1768 |
+
},
|
| 1769 |
+
{
|
| 1770 |
+
"epoch": 0.58,
|
| 1771 |
+
"learning_rate": 8.84294908575639e-05,
|
| 1772 |
+
"loss": 1.6116,
|
| 1773 |
+
"step": 290
|
| 1774 |
+
},
|
| 1775 |
+
{
|
| 1776 |
+
"epoch": 0.58,
|
| 1777 |
+
"learning_rate": 8.773168561218483e-05,
|
| 1778 |
+
"loss": 1.7391,
|
| 1779 |
+
"step": 291
|
| 1780 |
+
},
|
| 1781 |
+
{
|
| 1782 |
+
"epoch": 0.59,
|
| 1783 |
+
"learning_rate": 8.70344863600857e-05,
|
| 1784 |
+
"loss": 1.5991,
|
| 1785 |
+
"step": 292
|
| 1786 |
+
},
|
| 1787 |
+
{
|
| 1788 |
+
"epoch": 0.59,
|
| 1789 |
+
"learning_rate": 8.633792753941733e-05,
|
| 1790 |
+
"loss": 1.5397,
|
| 1791 |
+
"step": 293
|
| 1792 |
+
},
|
| 1793 |
+
{
|
| 1794 |
+
"epoch": 0.59,
|
| 1795 |
+
"learning_rate": 8.564204355669643e-05,
|
| 1796 |
+
"loss": 1.6073,
|
| 1797 |
+
"step": 294
|
| 1798 |
+
},
|
| 1799 |
+
{
|
| 1800 |
+
"epoch": 0.59,
|
| 1801 |
+
"learning_rate": 8.494686878510623e-05,
|
| 1802 |
+
"loss": 1.7106,
|
| 1803 |
+
"step": 295
|
| 1804 |
+
},
|
| 1805 |
+
{
|
| 1806 |
+
"epoch": 0.59,
|
| 1807 |
+
"learning_rate": 8.425243756279824e-05,
|
| 1808 |
+
"loss": 1.6159,
|
| 1809 |
+
"step": 296
|
| 1810 |
+
},
|
| 1811 |
+
{
|
| 1812 |
+
"epoch": 0.6,
|
| 1813 |
+
"learning_rate": 8.355878419119657e-05,
|
| 1814 |
+
"loss": 1.7041,
|
| 1815 |
+
"step": 297
|
| 1816 |
+
},
|
| 1817 |
+
{
|
| 1818 |
+
"epoch": 0.6,
|
| 1819 |
+
"learning_rate": 8.286594293330332e-05,
|
| 1820 |
+
"loss": 1.6621,
|
| 1821 |
+
"step": 298
|
| 1822 |
+
},
|
| 1823 |
+
{
|
| 1824 |
+
"epoch": 0.6,
|
| 1825 |
+
"learning_rate": 8.217394801200631e-05,
|
| 1826 |
+
"loss": 1.6209,
|
| 1827 |
+
"step": 299
|
| 1828 |
+
},
|
| 1829 |
+
{
|
| 1830 |
+
"epoch": 0.6,
|
| 1831 |
+
"learning_rate": 8.148283360838844e-05,
|
| 1832 |
+
"loss": 1.6256,
|
| 1833 |
+
"step": 300
|
| 1834 |
+
},
|
| 1835 |
+
{
|
| 1836 |
+
"epoch": 0.61,
|
| 1837 |
+
"learning_rate": 8.079263386003952e-05,
|
| 1838 |
+
"loss": 1.6747,
|
| 1839 |
+
"step": 301
|
| 1840 |
+
},
|
| 1841 |
+
{
|
| 1842 |
+
"epoch": 0.61,
|
| 1843 |
+
"learning_rate": 8.010338285937006e-05,
|
| 1844 |
+
"loss": 1.7257,
|
| 1845 |
+
"step": 302
|
| 1846 |
+
},
|
| 1847 |
+
{
|
| 1848 |
+
"epoch": 0.61,
|
| 1849 |
+
"learning_rate": 7.941511465192697e-05,
|
| 1850 |
+
"loss": 1.605,
|
| 1851 |
+
"step": 303
|
| 1852 |
+
},
|
| 1853 |
+
{
|
| 1854 |
+
"epoch": 0.61,
|
| 1855 |
+
"learning_rate": 7.872786323471232e-05,
|
| 1856 |
+
"loss": 1.5881,
|
| 1857 |
+
"step": 304
|
| 1858 |
+
},
|
| 1859 |
+
{
|
| 1860 |
+
"epoch": 0.61,
|
| 1861 |
+
"learning_rate": 7.804166255450373e-05,
|
| 1862 |
+
"loss": 1.6928,
|
| 1863 |
+
"step": 305
|
| 1864 |
+
},
|
| 1865 |
+
{
|
| 1866 |
+
"epoch": 0.62,
|
| 1867 |
+
"learning_rate": 7.735654650617763e-05,
|
| 1868 |
+
"loss": 1.6106,
|
| 1869 |
+
"step": 306
|
| 1870 |
+
},
|
| 1871 |
+
{
|
| 1872 |
+
"epoch": 0.62,
|
| 1873 |
+
"learning_rate": 7.667254893103519e-05,
|
| 1874 |
+
"loss": 1.6714,
|
| 1875 |
+
"step": 307
|
| 1876 |
+
},
|
| 1877 |
+
{
|
| 1878 |
+
"epoch": 0.62,
|
| 1879 |
+
"learning_rate": 7.598970361513051e-05,
|
| 1880 |
+
"loss": 1.5388,
|
| 1881 |
+
"step": 308
|
| 1882 |
+
},
|
| 1883 |
+
{
|
| 1884 |
+
"epoch": 0.62,
|
| 1885 |
+
"learning_rate": 7.53080442876019e-05,
|
| 1886 |
+
"loss": 1.546,
|
| 1887 |
+
"step": 309
|
| 1888 |
+
},
|
| 1889 |
+
{
|
| 1890 |
+
"epoch": 0.62,
|
| 1891 |
+
"learning_rate": 7.462760461900576e-05,
|
| 1892 |
+
"loss": 1.5758,
|
| 1893 |
+
"step": 310
|
| 1894 |
+
},
|
| 1895 |
+
{
|
| 1896 |
+
"epoch": 0.63,
|
| 1897 |
+
"learning_rate": 7.394841821965345e-05,
|
| 1898 |
+
"loss": 1.4756,
|
| 1899 |
+
"step": 311
|
| 1900 |
+
},
|
| 1901 |
+
{
|
| 1902 |
+
"epoch": 0.63,
|
| 1903 |
+
"learning_rate": 7.327051863795118e-05,
|
| 1904 |
+
"loss": 1.5601,
|
| 1905 |
+
"step": 312
|
| 1906 |
+
},
|
| 1907 |
+
{
|
| 1908 |
+
"epoch": 0.63,
|
| 1909 |
+
"learning_rate": 7.259393935874265e-05,
|
| 1910 |
+
"loss": 1.5874,
|
| 1911 |
+
"step": 313
|
| 1912 |
+
},
|
| 1913 |
+
{
|
| 1914 |
+
"epoch": 0.63,
|
| 1915 |
+
"learning_rate": 7.191871380165538e-05,
|
| 1916 |
+
"loss": 1.6352,
|
| 1917 |
+
"step": 314
|
| 1918 |
+
},
|
| 1919 |
+
{
|
| 1920 |
+
"epoch": 0.63,
|
| 1921 |
+
"learning_rate": 7.12448753194498e-05,
|
| 1922 |
+
"loss": 1.6191,
|
| 1923 |
+
"step": 315
|
| 1924 |
+
},
|
| 1925 |
+
{
|
| 1926 |
+
"epoch": 0.64,
|
| 1927 |
+
"learning_rate": 7.057245719637164e-05,
|
| 1928 |
+
"loss": 1.5642,
|
| 1929 |
+
"step": 316
|
| 1930 |
+
},
|
| 1931 |
+
{
|
| 1932 |
+
"epoch": 0.64,
|
| 1933 |
+
"learning_rate": 6.990149264650814e-05,
|
| 1934 |
+
"loss": 1.6193,
|
| 1935 |
+
"step": 317
|
| 1936 |
+
},
|
| 1937 |
+
{
|
| 1938 |
+
"epoch": 0.64,
|
| 1939 |
+
"learning_rate": 6.923201481214732e-05,
|
| 1940 |
+
"loss": 1.5273,
|
| 1941 |
+
"step": 318
|
| 1942 |
+
},
|
| 1943 |
+
{
|
| 1944 |
+
"epoch": 0.64,
|
| 1945 |
+
"learning_rate": 6.856405676214073e-05,
|
| 1946 |
+
"loss": 1.552,
|
| 1947 |
+
"step": 319
|
| 1948 |
+
},
|
| 1949 |
+
{
|
| 1950 |
+
"epoch": 0.64,
|
| 1951 |
+
"learning_rate": 6.789765149027039e-05,
|
| 1952 |
+
"loss": 1.4993,
|
| 1953 |
+
"step": 320
|
| 1954 |
+
},
|
| 1955 |
+
{
|
| 1956 |
+
"epoch": 0.65,
|
| 1957 |
+
"learning_rate": 6.723283191361873e-05,
|
| 1958 |
+
"loss": 1.5243,
|
| 1959 |
+
"step": 321
|
| 1960 |
+
},
|
| 1961 |
+
{
|
| 1962 |
+
"epoch": 0.65,
|
| 1963 |
+
"learning_rate": 6.656963087094284e-05,
|
| 1964 |
+
"loss": 1.6382,
|
| 1965 |
+
"step": 322
|
| 1966 |
+
},
|
| 1967 |
+
{
|
| 1968 |
+
"epoch": 0.65,
|
| 1969 |
+
"learning_rate": 6.590808112105232e-05,
|
| 1970 |
+
"loss": 1.6384,
|
| 1971 |
+
"step": 323
|
| 1972 |
+
},
|
| 1973 |
+
{
|
| 1974 |
+
"epoch": 0.65,
|
| 1975 |
+
"learning_rate": 6.524821534119114e-05,
|
| 1976 |
+
"loss": 1.5141,
|
| 1977 |
+
"step": 324
|
| 1978 |
+
},
|
| 1979 |
+
{
|
| 1980 |
+
"epoch": 0.65,
|
| 1981 |
+
"learning_rate": 6.459006612542365e-05,
|
| 1982 |
+
"loss": 1.526,
|
| 1983 |
+
"step": 325
|
| 1984 |
+
},
|
| 1985 |
+
{
|
| 1986 |
+
"epoch": 0.66,
|
| 1987 |
+
"learning_rate": 6.393366598302446e-05,
|
| 1988 |
+
"loss": 1.5187,
|
| 1989 |
+
"step": 326
|
| 1990 |
+
},
|
| 1991 |
+
{
|
| 1992 |
+
"epoch": 0.66,
|
| 1993 |
+
"learning_rate": 6.32790473368728e-05,
|
| 1994 |
+
"loss": 1.5672,
|
| 1995 |
+
"step": 327
|
| 1996 |
+
},
|
| 1997 |
+
{
|
| 1998 |
+
"epoch": 0.66,
|
| 1999 |
+
"learning_rate": 6.262624252185087e-05,
|
| 2000 |
+
"loss": 1.5509,
|
| 2001 |
+
"step": 328
|
| 2002 |
+
},
|
| 2003 |
+
{
|
| 2004 |
+
"epoch": 0.66,
|
| 2005 |
+
"learning_rate": 6.197528378324665e-05,
|
| 2006 |
+
"loss": 1.6236,
|
| 2007 |
+
"step": 329
|
| 2008 |
+
},
|
| 2009 |
+
{
|
| 2010 |
+
"epoch": 0.66,
|
| 2011 |
+
"learning_rate": 6.132620327516126e-05,
|
| 2012 |
+
"loss": 1.5135,
|
| 2013 |
+
"step": 330
|
| 2014 |
+
},
|
| 2015 |
+
{
|
| 2016 |
+
"epoch": 0.67,
|
| 2017 |
+
"learning_rate": 6.06790330589206e-05,
|
| 2018 |
+
"loss": 1.5453,
|
| 2019 |
+
"step": 331
|
| 2020 |
+
},
|
| 2021 |
+
{
|
| 2022 |
+
"epoch": 0.67,
|
| 2023 |
+
"learning_rate": 6.0033805101491794e-05,
|
| 2024 |
+
"loss": 1.4761,
|
| 2025 |
+
"step": 332
|
| 2026 |
+
},
|
| 2027 |
+
{
|
| 2028 |
+
"epoch": 0.67,
|
| 2029 |
+
"learning_rate": 5.939055127390396e-05,
|
| 2030 |
+
"loss": 1.5515,
|
| 2031 |
+
"step": 333
|
| 2032 |
+
},
|
| 2033 |
+
{
|
| 2034 |
+
"epoch": 0.67,
|
| 2035 |
+
"learning_rate": 5.8749303349674254e-05,
|
| 2036 |
+
"loss": 1.5779,
|
| 2037 |
+
"step": 334
|
| 2038 |
+
},
|
| 2039 |
+
{
|
| 2040 |
+
"epoch": 0.67,
|
| 2041 |
+
"learning_rate": 5.811009300323818e-05,
|
| 2042 |
+
"loss": 1.4634,
|
| 2043 |
+
"step": 335
|
| 2044 |
+
},
|
| 2045 |
+
{
|
| 2046 |
+
"epoch": 0.68,
|
| 2047 |
+
"learning_rate": 5.747295180838503e-05,
|
| 2048 |
+
"loss": 1.4856,
|
| 2049 |
+
"step": 336
|
| 2050 |
+
},
|
| 2051 |
+
{
|
| 2052 |
+
"epoch": 0.68,
|
| 2053 |
+
"learning_rate": 5.6837911236698536e-05,
|
| 2054 |
+
"loss": 1.4736,
|
| 2055 |
+
"step": 337
|
| 2056 |
+
},
|
| 2057 |
+
{
|
| 2058 |
+
"epoch": 0.68,
|
| 2059 |
+
"learning_rate": 5.620500265600206e-05,
|
| 2060 |
+
"loss": 1.4468,
|
| 2061 |
+
"step": 338
|
| 2062 |
+
},
|
| 2063 |
+
{
|
| 2064 |
+
"epoch": 0.68,
|
| 2065 |
+
"learning_rate": 5.5574257328809276e-05,
|
| 2066 |
+
"loss": 1.448,
|
| 2067 |
+
"step": 339
|
| 2068 |
+
},
|
| 2069 |
+
{
|
| 2070 |
+
"epoch": 0.68,
|
| 2071 |
+
"learning_rate": 5.494570641077999e-05,
|
| 2072 |
+
"loss": 1.5192,
|
| 2073 |
+
"step": 340
|
| 2074 |
+
},
|
| 2075 |
+
{
|
| 2076 |
+
"epoch": 0.69,
|
| 2077 |
+
"learning_rate": 5.431938094918132e-05,
|
| 2078 |
+
"loss": 1.4709,
|
| 2079 |
+
"step": 341
|
| 2080 |
+
},
|
| 2081 |
+
{
|
| 2082 |
+
"epoch": 0.69,
|
| 2083 |
+
"learning_rate": 5.369531188135368e-05,
|
| 2084 |
+
"loss": 1.4481,
|
| 2085 |
+
"step": 342
|
| 2086 |
+
},
|
| 2087 |
+
{
|
| 2088 |
+
"epoch": 0.69,
|
| 2089 |
+
"learning_rate": 5.307353003318325e-05,
|
| 2090 |
+
"loss": 1.5018,
|
| 2091 |
+
"step": 343
|
| 2092 |
+
},
|
| 2093 |
+
{
|
| 2094 |
+
"epoch": 0.69,
|
| 2095 |
+
"learning_rate": 5.2454066117578815e-05,
|
| 2096 |
+
"loss": 1.5606,
|
| 2097 |
+
"step": 344
|
| 2098 |
+
},
|
| 2099 |
+
{
|
| 2100 |
+
"epoch": 0.69,
|
| 2101 |
+
"learning_rate": 5.183695073295507e-05,
|
| 2102 |
+
"loss": 1.5407,
|
| 2103 |
+
"step": 345
|
| 2104 |
+
},
|
| 2105 |
+
{
|
| 2106 |
+
"epoch": 0.7,
|
| 2107 |
+
"learning_rate": 5.122221436172079e-05,
|
| 2108 |
+
"loss": 1.4878,
|
| 2109 |
+
"step": 346
|
| 2110 |
+
},
|
| 2111 |
+
{
|
| 2112 |
+
"epoch": 0.7,
|
| 2113 |
+
"learning_rate": 5.060988736877366e-05,
|
| 2114 |
+
"loss": 1.4904,
|
| 2115 |
+
"step": 347
|
| 2116 |
+
},
|
| 2117 |
+
{
|
| 2118 |
+
"epoch": 0.7,
|
| 2119 |
+
"learning_rate": 5.000000000000002e-05,
|
| 2120 |
+
"loss": 1.3881,
|
| 2121 |
+
"step": 348
|
| 2122 |
+
},
|
| 2123 |
+
{
|
| 2124 |
+
"epoch": 0.7,
|
| 2125 |
+
"learning_rate": 4.939258238078098e-05,
|
| 2126 |
+
"loss": 1.4511,
|
| 2127 |
+
"step": 349
|
| 2128 |
+
},
|
| 2129 |
+
{
|
| 2130 |
+
"epoch": 0.7,
|
| 2131 |
+
"learning_rate": 4.8787664514504504e-05,
|
| 2132 |
+
"loss": 1.4791,
|
| 2133 |
+
"step": 350
|
| 2134 |
+
},
|
| 2135 |
+
{
|
| 2136 |
+
"epoch": 0.71,
|
| 2137 |
+
"learning_rate": 4.818527628108317e-05,
|
| 2138 |
+
"loss": 1.323,
|
| 2139 |
+
"step": 351
|
| 2140 |
+
},
|
| 2141 |
+
{
|
| 2142 |
+
"epoch": 0.71,
|
| 2143 |
+
"learning_rate": 4.758544743547839e-05,
|
| 2144 |
+
"loss": 1.4073,
|
| 2145 |
+
"step": 352
|
| 2146 |
+
},
|
| 2147 |
+
{
|
| 2148 |
+
"epoch": 0.71,
|
| 2149 |
+
"learning_rate": 4.698820760623064e-05,
|
| 2150 |
+
"loss": 1.3715,
|
| 2151 |
+
"step": 353
|
| 2152 |
+
},
|
| 2153 |
+
{
|
| 2154 |
+
"epoch": 0.71,
|
| 2155 |
+
"learning_rate": 4.639358629399602e-05,
|
| 2156 |
+
"loss": 1.4781,
|
| 2157 |
+
"step": 354
|
| 2158 |
+
},
|
| 2159 |
+
{
|
| 2160 |
+
"epoch": 0.71,
|
| 2161 |
+
"learning_rate": 4.580161287008892e-05,
|
| 2162 |
+
"loss": 1.5291,
|
| 2163 |
+
"step": 355
|
| 2164 |
+
},
|
| 2165 |
+
{
|
| 2166 |
+
"epoch": 0.72,
|
| 2167 |
+
"learning_rate": 4.521231657503132e-05,
|
| 2168 |
+
"loss": 1.4415,
|
| 2169 |
+
"step": 356
|
| 2170 |
+
},
|
| 2171 |
+
{
|
| 2172 |
+
"epoch": 0.72,
|
| 2173 |
+
"learning_rate": 4.462572651710847e-05,
|
| 2174 |
+
"loss": 1.3697,
|
| 2175 |
+
"step": 357
|
| 2176 |
+
},
|
| 2177 |
+
{
|
| 2178 |
+
"epoch": 0.72,
|
| 2179 |
+
"learning_rate": 4.4041871670931135e-05,
|
| 2180 |
+
"loss": 1.5058,
|
| 2181 |
+
"step": 358
|
| 2182 |
+
},
|
| 2183 |
+
{
|
| 2184 |
+
"epoch": 0.72,
|
| 2185 |
+
"learning_rate": 4.346078087600412e-05,
|
| 2186 |
+
"loss": 1.3509,
|
| 2187 |
+
"step": 359
|
| 2188 |
+
},
|
| 2189 |
+
{
|
| 2190 |
+
"epoch": 0.72,
|
| 2191 |
+
"learning_rate": 4.288248283530214e-05,
|
| 2192 |
+
"loss": 1.5078,
|
| 2193 |
+
"step": 360
|
| 2194 |
+
},
|
| 2195 |
+
{
|
| 2196 |
+
"epoch": 0.73,
|
| 2197 |
+
"learning_rate": 4.230700611385174e-05,
|
| 2198 |
+
"loss": 1.4253,
|
| 2199 |
+
"step": 361
|
| 2200 |
+
},
|
| 2201 |
+
{
|
| 2202 |
+
"epoch": 0.73,
|
| 2203 |
+
"learning_rate": 4.173437913732048e-05,
|
| 2204 |
+
"loss": 1.3528,
|
| 2205 |
+
"step": 362
|
| 2206 |
+
},
|
| 2207 |
+
{
|
| 2208 |
+
"epoch": 0.73,
|
| 2209 |
+
"learning_rate": 4.116463019061269e-05,
|
| 2210 |
+
"loss": 1.4971,
|
| 2211 |
+
"step": 363
|
| 2212 |
+
},
|
| 2213 |
+
{
|
| 2214 |
+
"epoch": 0.73,
|
| 2215 |
+
"learning_rate": 4.059778741647261e-05,
|
| 2216 |
+
"loss": 1.4212,
|
| 2217 |
+
"step": 364
|
| 2218 |
+
},
|
| 2219 |
+
{
|
| 2220 |
+
"epoch": 0.73,
|
| 2221 |
+
"learning_rate": 4.003387881409397e-05,
|
| 2222 |
+
"loss": 1.4902,
|
| 2223 |
+
"step": 365
|
| 2224 |
+
},
|
| 2225 |
+
{
|
| 2226 |
+
"epoch": 0.74,
|
| 2227 |
+
"learning_rate": 3.947293223773715e-05,
|
| 2228 |
+
"loss": 1.4026,
|
| 2229 |
+
"step": 366
|
| 2230 |
+
},
|
| 2231 |
+
{
|
| 2232 |
+
"epoch": 0.74,
|
| 2233 |
+
"learning_rate": 3.8914975395353334e-05,
|
| 2234 |
+
"loss": 1.4036,
|
| 2235 |
+
"step": 367
|
| 2236 |
+
},
|
| 2237 |
+
{
|
| 2238 |
+
"epoch": 0.74,
|
| 2239 |
+
"learning_rate": 3.836003584721577e-05,
|
| 2240 |
+
"loss": 1.3422,
|
| 2241 |
+
"step": 368
|
| 2242 |
+
},
|
| 2243 |
+
{
|
| 2244 |
+
"epoch": 0.74,
|
| 2245 |
+
"learning_rate": 3.780814100455848e-05,
|
| 2246 |
+
"loss": 1.4615,
|
| 2247 |
+
"step": 369
|
| 2248 |
+
},
|
| 2249 |
+
{
|
| 2250 |
+
"epoch": 0.74,
|
| 2251 |
+
"learning_rate": 3.7259318128222276e-05,
|
| 2252 |
+
"loss": 1.4738,
|
| 2253 |
+
"step": 370
|
| 2254 |
+
},
|
| 2255 |
+
{
|
| 2256 |
+
"epoch": 0.75,
|
| 2257 |
+
"learning_rate": 3.671359432730834e-05,
|
| 2258 |
+
"loss": 1.2812,
|
| 2259 |
+
"step": 371
|
| 2260 |
+
},
|
| 2261 |
+
{
|
| 2262 |
+
"epoch": 0.75,
|
| 2263 |
+
"learning_rate": 3.617099655783884e-05,
|
| 2264 |
+
"loss": 1.3518,
|
| 2265 |
+
"step": 372
|
| 2266 |
+
},
|
| 2267 |
+
{
|
| 2268 |
+
"epoch": 0.75,
|
| 2269 |
+
"learning_rate": 3.563155162142584e-05,
|
| 2270 |
+
"loss": 1.3794,
|
| 2271 |
+
"step": 373
|
| 2272 |
+
},
|
| 2273 |
+
{
|
| 2274 |
+
"epoch": 0.75,
|
| 2275 |
+
"learning_rate": 3.509528616394716e-05,
|
| 2276 |
+
"loss": 1.344,
|
| 2277 |
+
"step": 374
|
| 2278 |
+
},
|
| 2279 |
+
{
|
| 2280 |
+
"epoch": 0.75,
|
| 2281 |
+
"learning_rate": 3.456222667423028e-05,
|
| 2282 |
+
"loss": 1.383,
|
| 2283 |
+
"step": 375
|
| 2284 |
+
},
|
| 2285 |
+
{
|
| 2286 |
+
"epoch": 0.75,
|
| 2287 |
+
"eval_loss": 1.3915969133377075,
|
| 2288 |
+
"eval_runtime": 52.0043,
|
| 2289 |
+
"eval_samples_per_second": 24.171,
|
| 2290 |
+
"eval_steps_per_second": 1.019,
|
| 2291 |
+
"step": 375
|
| 2292 |
+
},
|
| 2293 |
+
{
|
| 2294 |
+
"epoch": 0.76,
|
| 2295 |
+
"learning_rate": 3.403239948274392e-05,
|
| 2296 |
+
"loss": 1.3758,
|
| 2297 |
+
"step": 376
|
| 2298 |
+
},
|
| 2299 |
+
{
|
| 2300 |
+
"epoch": 0.76,
|
| 2301 |
+
"learning_rate": 3.350583076029754e-05,
|
| 2302 |
+
"loss": 1.3697,
|
| 2303 |
+
"step": 377
|
| 2304 |
+
},
|
| 2305 |
+
{
|
| 2306 |
+
"epoch": 0.76,
|
| 2307 |
+
"learning_rate": 3.298254651674848e-05,
|
| 2308 |
+
"loss": 1.3343,
|
| 2309 |
+
"step": 378
|
| 2310 |
+
},
|
| 2311 |
+
{
|
| 2312 |
+
"epoch": 0.76,
|
| 2313 |
+
"learning_rate": 3.246257259971727e-05,
|
| 2314 |
+
"loss": 1.4031,
|
| 2315 |
+
"step": 379
|
| 2316 |
+
},
|
| 2317 |
+
{
|
| 2318 |
+
"epoch": 0.76,
|
| 2319 |
+
"learning_rate": 3.1945934693310896e-05,
|
| 2320 |
+
"loss": 1.4939,
|
| 2321 |
+
"step": 380
|
| 2322 |
+
},
|
| 2323 |
+
{
|
| 2324 |
+
"epoch": 0.77,
|
| 2325 |
+
"learning_rate": 3.143265831685419e-05,
|
| 2326 |
+
"loss": 1.2987,
|
| 2327 |
+
"step": 381
|
| 2328 |
+
},
|
| 2329 |
+
{
|
| 2330 |
+
"epoch": 0.77,
|
| 2331 |
+
"learning_rate": 3.092276882362918e-05,
|
| 2332 |
+
"loss": 1.3531,
|
| 2333 |
+
"step": 382
|
| 2334 |
+
},
|
| 2335 |
+
{
|
| 2336 |
+
"epoch": 0.77,
|
| 2337 |
+
"learning_rate": 3.041629139962283e-05,
|
| 2338 |
+
"loss": 1.4016,
|
| 2339 |
+
"step": 383
|
| 2340 |
+
},
|
| 2341 |
+
{
|
| 2342 |
+
"epoch": 0.77,
|
| 2343 |
+
"learning_rate": 2.9913251062282986e-05,
|
| 2344 |
+
"loss": 1.4369,
|
| 2345 |
+
"step": 384
|
| 2346 |
+
},
|
| 2347 |
+
{
|
| 2348 |
+
"epoch": 0.77,
|
| 2349 |
+
"learning_rate": 2.9413672659282622e-05,
|
| 2350 |
+
"loss": 1.3164,
|
| 2351 |
+
"step": 385
|
| 2352 |
+
},
|
| 2353 |
+
{
|
| 2354 |
+
"epoch": 0.78,
|
| 2355 |
+
"learning_rate": 2.8917580867292526e-05,
|
| 2356 |
+
"loss": 1.4983,
|
| 2357 |
+
"step": 386
|
| 2358 |
+
},
|
| 2359 |
+
{
|
| 2360 |
+
"epoch": 0.78,
|
| 2361 |
+
"learning_rate": 2.8425000190762353e-05,
|
| 2362 |
+
"loss": 1.2823,
|
| 2363 |
+
"step": 387
|
| 2364 |
+
},
|
| 2365 |
+
{
|
| 2366 |
+
"epoch": 0.78,
|
| 2367 |
+
"learning_rate": 2.793595496071021e-05,
|
| 2368 |
+
"loss": 1.3993,
|
| 2369 |
+
"step": 388
|
| 2370 |
+
},
|
| 2371 |
+
{
|
| 2372 |
+
"epoch": 0.78,
|
| 2373 |
+
"learning_rate": 2.7450469333520855e-05,
|
| 2374 |
+
"loss": 1.3771,
|
| 2375 |
+
"step": 389
|
| 2376 |
+
},
|
| 2377 |
+
{
|
| 2378 |
+
"epoch": 0.78,
|
| 2379 |
+
"learning_rate": 2.6968567289752578e-05,
|
| 2380 |
+
"loss": 1.3688,
|
| 2381 |
+
"step": 390
|
| 2382 |
+
},
|
| 2383 |
+
{
|
| 2384 |
+
"epoch": 0.79,
|
| 2385 |
+
"learning_rate": 2.6490272632952505e-05,
|
| 2386 |
+
"loss": 1.3941,
|
| 2387 |
+
"step": 391
|
| 2388 |
+
},
|
| 2389 |
+
{
|
| 2390 |
+
"epoch": 0.79,
|
| 2391 |
+
"learning_rate": 2.6015608988480955e-05,
|
| 2392 |
+
"loss": 1.4538,
|
| 2393 |
+
"step": 392
|
| 2394 |
+
},
|
| 2395 |
+
{
|
| 2396 |
+
"epoch": 0.79,
|
| 2397 |
+
"learning_rate": 2.5544599802344394e-05,
|
| 2398 |
+
"loss": 1.3141,
|
| 2399 |
+
"step": 393
|
| 2400 |
+
},
|
| 2401 |
+
{
|
| 2402 |
+
"epoch": 0.79,
|
| 2403 |
+
"learning_rate": 2.5077268340037454e-05,
|
| 2404 |
+
"loss": 1.4652,
|
| 2405 |
+
"step": 394
|
| 2406 |
+
},
|
| 2407 |
+
{
|
| 2408 |
+
"epoch": 0.79,
|
| 2409 |
+
"learning_rate": 2.4613637685393432e-05,
|
| 2410 |
+
"loss": 1.2551,
|
| 2411 |
+
"step": 395
|
| 2412 |
+
},
|
| 2413 |
+
{
|
| 2414 |
+
"epoch": 0.8,
|
| 2415 |
+
"learning_rate": 2.415373073944449e-05,
|
| 2416 |
+
"loss": 1.416,
|
| 2417 |
+
"step": 396
|
| 2418 |
+
},
|
| 2419 |
+
{
|
| 2420 |
+
"epoch": 0.8,
|
| 2421 |
+
"learning_rate": 2.3697570219290077e-05,
|
| 2422 |
+
"loss": 1.351,
|
| 2423 |
+
"step": 397
|
| 2424 |
+
},
|
| 2425 |
+
{
|
| 2426 |
+
"epoch": 0.8,
|
| 2427 |
+
"learning_rate": 2.324517865697501e-05,
|
| 2428 |
+
"loss": 1.3533,
|
| 2429 |
+
"step": 398
|
| 2430 |
+
},
|
| 2431 |
+
{
|
| 2432 |
+
"epoch": 0.8,
|
| 2433 |
+
"learning_rate": 2.2796578398376523e-05,
|
| 2434 |
+
"loss": 1.2773,
|
| 2435 |
+
"step": 399
|
| 2436 |
+
},
|
| 2437 |
+
{
|
| 2438 |
+
"epoch": 0.8,
|
| 2439 |
+
"learning_rate": 2.235179160210037e-05,
|
| 2440 |
+
"loss": 1.3286,
|
| 2441 |
+
"step": 400
|
| 2442 |
+
},
|
| 2443 |
+
{
|
| 2444 |
+
"epoch": 0.81,
|
| 2445 |
+
"learning_rate": 2.1910840238386398e-05,
|
| 2446 |
+
"loss": 1.3392,
|
| 2447 |
+
"step": 401
|
| 2448 |
+
},
|
| 2449 |
+
{
|
| 2450 |
+
"epoch": 0.81,
|
| 2451 |
+
"learning_rate": 2.147374608802326e-05,
|
| 2452 |
+
"loss": 1.2921,
|
| 2453 |
+
"step": 402
|
| 2454 |
+
},
|
| 2455 |
+
{
|
| 2456 |
+
"epoch": 0.81,
|
| 2457 |
+
"learning_rate": 2.104053074127268e-05,
|
| 2458 |
+
"loss": 1.3551,
|
| 2459 |
+
"step": 403
|
| 2460 |
+
},
|
| 2461 |
+
{
|
| 2462 |
+
"epoch": 0.81,
|
| 2463 |
+
"learning_rate": 2.06112155968028e-05,
|
| 2464 |
+
"loss": 1.2801,
|
| 2465 |
+
"step": 404
|
| 2466 |
+
},
|
| 2467 |
+
{
|
| 2468 |
+
"epoch": 0.81,
|
| 2469 |
+
"learning_rate": 2.0185821860631394e-05,
|
| 2470 |
+
"loss": 1.3758,
|
| 2471 |
+
"step": 405
|
| 2472 |
+
},
|
| 2473 |
+
{
|
| 2474 |
+
"epoch": 0.82,
|
| 2475 |
+
"learning_rate": 1.9764370545078215e-05,
|
| 2476 |
+
"loss": 1.3692,
|
| 2477 |
+
"step": 406
|
| 2478 |
+
},
|
| 2479 |
+
{
|
| 2480 |
+
"epoch": 0.82,
|
| 2481 |
+
"learning_rate": 1.9346882467727325e-05,
|
| 2482 |
+
"loss": 1.2987,
|
| 2483 |
+
"step": 407
|
| 2484 |
+
},
|
| 2485 |
+
{
|
| 2486 |
+
"epoch": 0.82,
|
| 2487 |
+
"learning_rate": 1.893337825039849e-05,
|
| 2488 |
+
"loss": 1.2623,
|
| 2489 |
+
"step": 408
|
| 2490 |
+
},
|
| 2491 |
+
{
|
| 2492 |
+
"epoch": 0.82,
|
| 2493 |
+
"learning_rate": 1.852387831812893e-05,
|
| 2494 |
+
"loss": 1.4118,
|
| 2495 |
+
"step": 409
|
| 2496 |
+
},
|
| 2497 |
+
{
|
| 2498 |
+
"epoch": 0.82,
|
| 2499 |
+
"learning_rate": 1.811840289816409e-05,
|
| 2500 |
+
"loss": 1.3289,
|
| 2501 |
+
"step": 410
|
| 2502 |
+
},
|
| 2503 |
+
{
|
| 2504 |
+
"epoch": 0.83,
|
| 2505 |
+
"learning_rate": 1.7716972018958766e-05,
|
| 2506 |
+
"loss": 1.37,
|
| 2507 |
+
"step": 411
|
| 2508 |
+
},
|
| 2509 |
+
{
|
| 2510 |
+
"epoch": 0.83,
|
| 2511 |
+
"learning_rate": 1.7319605509187608e-05,
|
| 2512 |
+
"loss": 1.3297,
|
| 2513 |
+
"step": 412
|
| 2514 |
+
},
|
| 2515 |
+
{
|
| 2516 |
+
"epoch": 0.83,
|
| 2517 |
+
"learning_rate": 1.6926322996765897e-05,
|
| 2518 |
+
"loss": 1.3106,
|
| 2519 |
+
"step": 413
|
| 2520 |
+
},
|
| 2521 |
+
{
|
| 2522 |
+
"epoch": 0.83,
|
| 2523 |
+
"learning_rate": 1.653714390787979e-05,
|
| 2524 |
+
"loss": 1.2543,
|
| 2525 |
+
"step": 414
|
| 2526 |
+
},
|
| 2527 |
+
{
|
| 2528 |
+
"epoch": 0.83,
|
| 2529 |
+
"learning_rate": 1.61520874660269e-05,
|
| 2530 |
+
"loss": 1.1748,
|
| 2531 |
+
"step": 415
|
| 2532 |
+
},
|
| 2533 |
+
{
|
| 2534 |
+
"epoch": 0.84,
|
| 2535 |
+
"learning_rate": 1.5771172691066794e-05,
|
| 2536 |
+
"loss": 1.3091,
|
| 2537 |
+
"step": 416
|
| 2538 |
+
},
|
| 2539 |
+
{
|
| 2540 |
+
"epoch": 0.84,
|
| 2541 |
+
"learning_rate": 1.5394418398281352e-05,
|
| 2542 |
+
"loss": 1.3359,
|
| 2543 |
+
"step": 417
|
| 2544 |
+
},
|
| 2545 |
+
{
|
| 2546 |
+
"epoch": 0.84,
|
| 2547 |
+
"learning_rate": 1.5021843197445528e-05,
|
| 2548 |
+
"loss": 1.361,
|
| 2549 |
+
"step": 418
|
| 2550 |
+
},
|
| 2551 |
+
{
|
| 2552 |
+
"epoch": 0.84,
|
| 2553 |
+
"learning_rate": 1.4653465491908003e-05,
|
| 2554 |
+
"loss": 1.3599,
|
| 2555 |
+
"step": 419
|
| 2556 |
+
},
|
| 2557 |
+
{
|
| 2558 |
+
"epoch": 0.84,
|
| 2559 |
+
"learning_rate": 1.4289303477682347e-05,
|
| 2560 |
+
"loss": 1.3696,
|
| 2561 |
+
"step": 420
|
| 2562 |
+
},
|
| 2563 |
+
{
|
| 2564 |
+
"epoch": 0.85,
|
| 2565 |
+
"learning_rate": 1.3929375142547917e-05,
|
| 2566 |
+
"loss": 1.1577,
|
| 2567 |
+
"step": 421
|
| 2568 |
+
},
|
| 2569 |
+
{
|
| 2570 |
+
"epoch": 0.85,
|
| 2571 |
+
"learning_rate": 1.3573698265161683e-05,
|
| 2572 |
+
"loss": 1.3131,
|
| 2573 |
+
"step": 422
|
| 2574 |
+
},
|
| 2575 |
+
{
|
| 2576 |
+
"epoch": 0.85,
|
| 2577 |
+
"learning_rate": 1.3222290414179794e-05,
|
| 2578 |
+
"loss": 1.2763,
|
| 2579 |
+
"step": 423
|
| 2580 |
+
},
|
| 2581 |
+
{
|
| 2582 |
+
"epoch": 0.85,
|
| 2583 |
+
"learning_rate": 1.2875168947389981e-05,
|
| 2584 |
+
"loss": 1.2822,
|
| 2585 |
+
"step": 424
|
| 2586 |
+
},
|
| 2587 |
+
{
|
| 2588 |
+
"epoch": 0.85,
|
| 2589 |
+
"learning_rate": 1.2532351010853916e-05,
|
| 2590 |
+
"loss": 1.3403,
|
| 2591 |
+
"step": 425
|
| 2592 |
+
},
|
| 2593 |
+
{
|
| 2594 |
+
"epoch": 0.86,
|
| 2595 |
+
"learning_rate": 1.2193853538060595e-05,
|
| 2596 |
+
"loss": 1.2745,
|
| 2597 |
+
"step": 426
|
| 2598 |
+
},
|
| 2599 |
+
{
|
| 2600 |
+
"epoch": 0.86,
|
| 2601 |
+
"learning_rate": 1.1859693249089642e-05,
|
| 2602 |
+
"loss": 1.3022,
|
| 2603 |
+
"step": 427
|
| 2604 |
+
},
|
| 2605 |
+
{
|
| 2606 |
+
"epoch": 0.86,
|
| 2607 |
+
"learning_rate": 1.152988664978556e-05,
|
| 2608 |
+
"loss": 1.3088,
|
| 2609 |
+
"step": 428
|
| 2610 |
+
},
|
| 2611 |
+
{
|
| 2612 |
+
"epoch": 0.86,
|
| 2613 |
+
"learning_rate": 1.1204450030942348e-05,
|
| 2614 |
+
"loss": 1.3025,
|
| 2615 |
+
"step": 429
|
| 2616 |
+
},
|
| 2617 |
+
{
|
| 2618 |
+
"epoch": 0.86,
|
| 2619 |
+
"learning_rate": 1.0883399467498956e-05,
|
| 2620 |
+
"loss": 1.3092,
|
| 2621 |
+
"step": 430
|
| 2622 |
+
},
|
| 2623 |
+
{
|
| 2624 |
+
"epoch": 0.87,
|
| 2625 |
+
"learning_rate": 1.0566750817745074e-05,
|
| 2626 |
+
"loss": 1.3048,
|
| 2627 |
+
"step": 431
|
| 2628 |
+
},
|
| 2629 |
+
{
|
| 2630 |
+
"epoch": 0.87,
|
| 2631 |
+
"learning_rate": 1.0254519722537947e-05,
|
| 2632 |
+
"loss": 1.2341,
|
| 2633 |
+
"step": 432
|
| 2634 |
+
},
|
| 2635 |
+
{
|
| 2636 |
+
"epoch": 0.87,
|
| 2637 |
+
"learning_rate": 9.946721604529718e-06,
|
| 2638 |
+
"loss": 1.3788,
|
| 2639 |
+
"step": 433
|
| 2640 |
+
},
|
| 2641 |
+
{
|
| 2642 |
+
"epoch": 0.87,
|
| 2643 |
+
"learning_rate": 9.643371667405698e-06,
|
| 2644 |
+
"loss": 1.4034,
|
| 2645 |
+
"step": 434
|
| 2646 |
+
},
|
| 2647 |
+
{
|
| 2648 |
+
"epoch": 0.87,
|
| 2649 |
+
"learning_rate": 9.344484895133342e-06,
|
| 2650 |
+
"loss": 1.3039,
|
| 2651 |
+
"step": 435
|
| 2652 |
+
},
|
| 2653 |
+
{
|
| 2654 |
+
"epoch": 0.88,
|
| 2655 |
+
"learning_rate": 9.050076051222067e-06,
|
| 2656 |
+
"loss": 1.3125,
|
| 2657 |
+
"step": 436
|
| 2658 |
+
},
|
| 2659 |
+
{
|
| 2660 |
+
"epoch": 0.88,
|
| 2661 |
+
"learning_rate": 8.760159677994172e-06,
|
| 2662 |
+
"loss": 1.3098,
|
| 2663 |
+
"step": 437
|
| 2664 |
+
},
|
| 2665 |
+
{
|
| 2666 |
+
"epoch": 0.88,
|
| 2667 |
+
"learning_rate": 8.474750095866236e-06,
|
| 2668 |
+
"loss": 1.2353,
|
| 2669 |
+
"step": 438
|
| 2670 |
+
},
|
| 2671 |
+
{
|
| 2672 |
+
"epoch": 0.88,
|
| 2673 |
+
"learning_rate": 8.193861402642088e-06,
|
| 2674 |
+
"loss": 1.3627,
|
| 2675 |
+
"step": 439
|
| 2676 |
+
},
|
| 2677 |
+
{
|
| 2678 |
+
"epoch": 0.88,
|
| 2679 |
+
"learning_rate": 7.91750747281621e-06,
|
| 2680 |
+
"loss": 1.4489,
|
| 2681 |
+
"step": 440
|
| 2682 |
+
},
|
| 2683 |
+
{
|
| 2684 |
+
"epoch": 0.89,
|
| 2685 |
+
"learning_rate": 7.645701956888507e-06,
|
| 2686 |
+
"loss": 1.3553,
|
| 2687 |
+
"step": 441
|
| 2688 |
+
},
|
| 2689 |
+
{
|
| 2690 |
+
"epoch": 0.89,
|
| 2691 |
+
"learning_rate": 7.378458280689993e-06,
|
| 2692 |
+
"loss": 1.3337,
|
| 2693 |
+
"step": 442
|
| 2694 |
+
},
|
| 2695 |
+
{
|
| 2696 |
+
"epoch": 0.89,
|
| 2697 |
+
"learning_rate": 7.115789644719728e-06,
|
| 2698 |
+
"loss": 1.2856,
|
| 2699 |
+
"step": 443
|
| 2700 |
+
},
|
| 2701 |
+
{
|
| 2702 |
+
"epoch": 0.89,
|
| 2703 |
+
"learning_rate": 6.857709023492587e-06,
|
| 2704 |
+
"loss": 1.3764,
|
| 2705 |
+
"step": 444
|
| 2706 |
+
},
|
| 2707 |
+
{
|
| 2708 |
+
"epoch": 0.89,
|
| 2709 |
+
"learning_rate": 6.60422916489859e-06,
|
| 2710 |
+
"loss": 1.2939,
|
| 2711 |
+
"step": 445
|
| 2712 |
+
},
|
| 2713 |
+
{
|
| 2714 |
+
"epoch": 0.9,
|
| 2715 |
+
"learning_rate": 6.355362589573077e-06,
|
| 2716 |
+
"loss": 1.3477,
|
| 2717 |
+
"step": 446
|
| 2718 |
+
},
|
| 2719 |
+
{
|
| 2720 |
+
"epoch": 0.9,
|
| 2721 |
+
"learning_rate": 6.111121590278346e-06,
|
| 2722 |
+
"loss": 1.3298,
|
| 2723 |
+
"step": 447
|
| 2724 |
+
},
|
| 2725 |
+
{
|
| 2726 |
+
"epoch": 0.9,
|
| 2727 |
+
"learning_rate": 5.8715182312963575e-06,
|
| 2728 |
+
"loss": 1.3331,
|
| 2729 |
+
"step": 448
|
| 2730 |
+
},
|
| 2731 |
+
{
|
| 2732 |
+
"epoch": 0.9,
|
| 2733 |
+
"learning_rate": 5.636564347832907e-06,
|
| 2734 |
+
"loss": 1.2503,
|
| 2735 |
+
"step": 449
|
| 2736 |
+
},
|
| 2737 |
+
{
|
| 2738 |
+
"epoch": 0.9,
|
| 2739 |
+
"learning_rate": 5.4062715454329726e-06,
|
| 2740 |
+
"loss": 1.2786,
|
| 2741 |
+
"step": 450
|
| 2742 |
+
},
|
| 2743 |
+
{
|
| 2744 |
+
"epoch": 0.91,
|
| 2745 |
+
"learning_rate": 5.180651199407449e-06,
|
| 2746 |
+
"loss": 1.286,
|
| 2747 |
+
"step": 451
|
| 2748 |
+
},
|
| 2749 |
+
{
|
| 2750 |
+
"epoch": 0.91,
|
| 2751 |
+
"learning_rate": 4.959714454271369e-06,
|
| 2752 |
+
"loss": 1.2775,
|
| 2753 |
+
"step": 452
|
| 2754 |
+
},
|
| 2755 |
+
{
|
| 2756 |
+
"epoch": 0.91,
|
| 2757 |
+
"learning_rate": 4.7434722231932685e-06,
|
| 2758 |
+
"loss": 1.2779,
|
| 2759 |
+
"step": 453
|
| 2760 |
+
},
|
| 2761 |
+
{
|
| 2762 |
+
"epoch": 0.91,
|
| 2763 |
+
"learning_rate": 4.531935187456216e-06,
|
| 2764 |
+
"loss": 1.315,
|
| 2765 |
+
"step": 454
|
| 2766 |
+
},
|
| 2767 |
+
{
|
| 2768 |
+
"epoch": 0.91,
|
| 2769 |
+
"learning_rate": 4.325113795930203e-06,
|
| 2770 |
+
"loss": 1.3792,
|
| 2771 |
+
"step": 455
|
| 2772 |
+
},
|
| 2773 |
+
{
|
| 2774 |
+
"epoch": 0.92,
|
| 2775 |
+
"learning_rate": 4.1230182645560555e-06,
|
| 2776 |
+
"loss": 1.4314,
|
| 2777 |
+
"step": 456
|
| 2778 |
+
},
|
| 2779 |
+
{
|
| 2780 |
+
"epoch": 0.92,
|
| 2781 |
+
"learning_rate": 3.925658575840696e-06,
|
| 2782 |
+
"loss": 1.2617,
|
| 2783 |
+
"step": 457
|
| 2784 |
+
},
|
| 2785 |
+
{
|
| 2786 |
+
"epoch": 0.92,
|
| 2787 |
+
"learning_rate": 3.7330444783642338e-06,
|
| 2788 |
+
"loss": 1.3592,
|
| 2789 |
+
"step": 458
|
| 2790 |
+
},
|
| 2791 |
+
{
|
| 2792 |
+
"epoch": 0.92,
|
| 2793 |
+
"learning_rate": 3.5451854862982746e-06,
|
| 2794 |
+
"loss": 1.3299,
|
| 2795 |
+
"step": 459
|
| 2796 |
+
},
|
| 2797 |
+
{
|
| 2798 |
+
"epoch": 0.92,
|
| 2799 |
+
"learning_rate": 3.3620908789360863e-06,
|
| 2800 |
+
"loss": 1.3729,
|
| 2801 |
+
"step": 460
|
| 2802 |
+
},
|
| 2803 |
+
{
|
| 2804 |
+
"epoch": 0.93,
|
| 2805 |
+
"learning_rate": 3.1837697002341293e-06,
|
| 2806 |
+
"loss": 1.2571,
|
| 2807 |
+
"step": 461
|
| 2808 |
+
},
|
| 2809 |
+
{
|
| 2810 |
+
"epoch": 0.93,
|
| 2811 |
+
"learning_rate": 3.010230758365462e-06,
|
| 2812 |
+
"loss": 1.2179,
|
| 2813 |
+
"step": 462
|
| 2814 |
+
},
|
| 2815 |
+
{
|
| 2816 |
+
"epoch": 0.93,
|
| 2817 |
+
"learning_rate": 2.841482625284564e-06,
|
| 2818 |
+
"loss": 1.2873,
|
| 2819 |
+
"step": 463
|
| 2820 |
+
},
|
| 2821 |
+
{
|
| 2822 |
+
"epoch": 0.93,
|
| 2823 |
+
"learning_rate": 2.677533636303964e-06,
|
| 2824 |
+
"loss": 1.3169,
|
| 2825 |
+
"step": 464
|
| 2826 |
+
},
|
| 2827 |
+
{
|
| 2828 |
+
"epoch": 0.93,
|
| 2829 |
+
"learning_rate": 2.518391889682525e-06,
|
| 2830 |
+
"loss": 1.2881,
|
| 2831 |
+
"step": 465
|
| 2832 |
+
},
|
| 2833 |
+
{
|
| 2834 |
+
"epoch": 0.94,
|
| 2835 |
+
"learning_rate": 2.3640652462253886e-06,
|
| 2836 |
+
"loss": 1.2479,
|
| 2837 |
+
"step": 466
|
| 2838 |
+
},
|
| 2839 |
+
{
|
| 2840 |
+
"epoch": 0.94,
|
| 2841 |
+
"learning_rate": 2.2145613288957478e-06,
|
| 2842 |
+
"loss": 1.2183,
|
| 2843 |
+
"step": 467
|
| 2844 |
+
},
|
| 2845 |
+
{
|
| 2846 |
+
"epoch": 0.94,
|
| 2847 |
+
"learning_rate": 2.069887522438252e-06,
|
| 2848 |
+
"loss": 1.3322,
|
| 2849 |
+
"step": 468
|
| 2850 |
+
},
|
| 2851 |
+
{
|
| 2852 |
+
"epoch": 0.94,
|
| 2853 |
+
"learning_rate": 1.9300509730142855e-06,
|
| 2854 |
+
"loss": 1.2666,
|
| 2855 |
+
"step": 469
|
| 2856 |
+
},
|
| 2857 |
+
{
|
| 2858 |
+
"epoch": 0.94,
|
| 2859 |
+
"learning_rate": 1.7950585878489856e-06,
|
| 2860 |
+
"loss": 1.2443,
|
| 2861 |
+
"step": 470
|
| 2862 |
+
},
|
| 2863 |
+
{
|
| 2864 |
+
"epoch": 0.95,
|
| 2865 |
+
"learning_rate": 1.6649170348899789e-06,
|
| 2866 |
+
"loss": 1.2989,
|
| 2867 |
+
"step": 471
|
| 2868 |
+
},
|
| 2869 |
+
{
|
| 2870 |
+
"epoch": 0.95,
|
| 2871 |
+
"learning_rate": 1.5396327424781366e-06,
|
| 2872 |
+
"loss": 1.2325,
|
| 2873 |
+
"step": 472
|
| 2874 |
+
},
|
| 2875 |
+
{
|
| 2876 |
+
"epoch": 0.95,
|
| 2877 |
+
"learning_rate": 1.4192118990299707e-06,
|
| 2878 |
+
"loss": 1.3192,
|
| 2879 |
+
"step": 473
|
| 2880 |
+
},
|
| 2881 |
+
{
|
| 2882 |
+
"epoch": 0.95,
|
| 2883 |
+
"learning_rate": 1.3036604527319474e-06,
|
| 2884 |
+
"loss": 1.2725,
|
| 2885 |
+
"step": 474
|
| 2886 |
+
},
|
| 2887 |
+
{
|
| 2888 |
+
"epoch": 0.95,
|
| 2889 |
+
"learning_rate": 1.1929841112467533e-06,
|
| 2890 |
+
"loss": 1.3761,
|
| 2891 |
+
"step": 475
|
| 2892 |
+
},
|
| 2893 |
+
{
|
| 2894 |
+
"epoch": 0.96,
|
| 2895 |
+
"learning_rate": 1.0871883414312777e-06,
|
| 2896 |
+
"loss": 1.3084,
|
| 2897 |
+
"step": 476
|
| 2898 |
+
},
|
| 2899 |
+
{
|
| 2900 |
+
"epoch": 0.96,
|
| 2901 |
+
"learning_rate": 9.862783690666178e-07,
|
| 2902 |
+
"loss": 1.2498,
|
| 2903 |
+
"step": 477
|
| 2904 |
+
},
|
| 2905 |
+
{
|
| 2906 |
+
"epoch": 0.96,
|
| 2907 |
+
"learning_rate": 8.902591785999725e-07,
|
| 2908 |
+
"loss": 1.3384,
|
| 2909 |
+
"step": 478
|
| 2910 |
+
},
|
| 2911 |
+
{
|
| 2912 |
+
"epoch": 0.96,
|
| 2913 |
+
"learning_rate": 7.991355128984079e-07,
|
| 2914 |
+
"loss": 1.2835,
|
| 2915 |
+
"step": 479
|
| 2916 |
+
},
|
| 2917 |
+
{
|
| 2918 |
+
"epoch": 0.96,
|
| 2919 |
+
"learning_rate": 7.129118730145656e-07,
|
| 2920 |
+
"loss": 1.412,
|
| 2921 |
+
"step": 480
|
| 2922 |
+
},
|
| 2923 |
+
{
|
| 2924 |
+
"epoch": 0.97,
|
| 2925 |
+
"learning_rate": 6.315925179643744e-07,
|
| 2926 |
+
"loss": 1.2562,
|
| 2927 |
+
"step": 481
|
| 2928 |
+
},
|
| 2929 |
+
{
|
| 2930 |
+
"epoch": 0.97,
|
| 2931 |
+
"learning_rate": 5.55181464516652e-07,
|
| 2932 |
+
"loss": 1.3095,
|
| 2933 |
+
"step": 482
|
| 2934 |
+
},
|
| 2935 |
+
{
|
| 2936 |
+
"epoch": 0.97,
|
| 2937 |
+
"learning_rate": 4.836824869946965e-07,
|
| 2938 |
+
"loss": 1.252,
|
| 2939 |
+
"step": 483
|
| 2940 |
+
},
|
| 2941 |
+
{
|
| 2942 |
+
"epoch": 0.97,
|
| 2943 |
+
"learning_rate": 4.170991170898808e-07,
|
| 2944 |
+
"loss": 1.3766,
|
| 2945 |
+
"step": 484
|
| 2946 |
+
},
|
| 2947 |
+
{
|
| 2948 |
+
"epoch": 0.97,
|
| 2949 |
+
"learning_rate": 3.554346436871581e-07,
|
| 2950 |
+
"loss": 1.2876,
|
| 2951 |
+
"step": 485
|
| 2952 |
+
},
|
| 2953 |
+
{
|
| 2954 |
+
"epoch": 0.98,
|
| 2955 |
+
"learning_rate": 2.986921127026476e-07,
|
| 2956 |
+
"loss": 1.3465,
|
| 2957 |
+
"step": 486
|
| 2958 |
+
},
|
| 2959 |
+
{
|
| 2960 |
+
"epoch": 0.98,
|
| 2961 |
+
"learning_rate": 2.468743269331442e-07,
|
| 2962 |
+
"loss": 1.3842,
|
| 2963 |
+
"step": 487
|
| 2964 |
+
},
|
| 2965 |
+
{
|
| 2966 |
+
"epoch": 0.98,
|
| 2967 |
+
"learning_rate": 1.9998384591773944e-07,
|
| 2968 |
+
"loss": 1.3558,
|
| 2969 |
+
"step": 488
|
| 2970 |
+
},
|
| 2971 |
+
{
|
| 2972 |
+
"epoch": 0.98,
|
| 2973 |
+
"learning_rate": 1.5802298581132358e-07,
|
| 2974 |
+
"loss": 1.2995,
|
| 2975 |
+
"step": 489
|
| 2976 |
+
},
|
| 2977 |
+
{
|
| 2978 |
+
"epoch": 0.98,
|
| 2979 |
+
"learning_rate": 1.209938192701876e-07,
|
| 2980 |
+
"loss": 1.27,
|
| 2981 |
+
"step": 490
|
| 2982 |
+
},
|
| 2983 |
+
{
|
| 2984 |
+
"epoch": 0.99,
|
| 2985 |
+
"learning_rate": 8.889817534969425e-08,
|
| 2986 |
+
"loss": 1.2425,
|
| 2987 |
+
"step": 491
|
| 2988 |
+
},
|
| 2989 |
+
{
|
| 2990 |
+
"epoch": 0.99,
|
| 2991 |
+
"learning_rate": 6.173763941389465e-08,
|
| 2992 |
+
"loss": 1.2134,
|
| 2993 |
+
"step": 492
|
| 2994 |
+
},
|
| 2995 |
+
{
|
| 2996 |
+
"epoch": 0.99,
|
| 2997 |
+
"learning_rate": 3.9513553057202165e-08,
|
| 2998 |
+
"loss": 1.2438,
|
| 2999 |
+
"step": 493
|
| 3000 |
+
},
|
| 3001 |
+
{
|
| 3002 |
+
"epoch": 0.99,
|
| 3003 |
+
"learning_rate": 2.222701403818972e-08,
|
| 3004 |
+
"loss": 1.329,
|
| 3005 |
+
"step": 494
|
| 3006 |
+
},
|
| 3007 |
+
{
|
| 3008 |
+
"epoch": 0.99,
|
| 3009 |
+
"learning_rate": 9.878876225277722e-09,
|
| 3010 |
+
"loss": 1.4423,
|
| 3011 |
+
"step": 495
|
| 3012 |
+
},
|
| 3013 |
+
{
|
| 3014 |
+
"epoch": 1.0,
|
| 3015 |
+
"learning_rate": 2.469749554634415e-09,
|
| 3016 |
+
"loss": 1.2607,
|
| 3017 |
+
"step": 496
|
| 3018 |
+
},
|
| 3019 |
+
{
|
| 3020 |
+
"epoch": 1.0,
|
| 3021 |
+
"learning_rate": 0.0,
|
| 3022 |
+
"loss": 1.3635,
|
| 3023 |
+
"step": 497
|
| 3024 |
+
}
|
| 3025 |
+
],
|
| 3026 |
+
"logging_steps": 1,
|
| 3027 |
+
"max_steps": 497,
|
| 3028 |
+
"num_input_tokens_seen": 0,
|
| 3029 |
+
"num_train_epochs": 1,
|
| 3030 |
+
"save_steps": 500,
|
| 3031 |
+
"total_flos": 3.751894162656461e+17,
|
| 3032 |
+
"train_batch_size": 24,
|
| 3033 |
+
"trial_name": null,
|
| 3034 |
+
"trial_params": null
|
| 3035 |
+
}
|
checkpoint-497/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c4cd1b91f73c711df87755d95aedeff9f4efb28c7ade5b5665d8c0eadb89c248
|
| 3 |
+
size 5304
|
pytorch_model.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 2690915926
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:430773bcc382aa85105ac14dfbe701477264c4471e671bcfd04fbedb29a7d17d
|
| 3 |
size 2690915926
|
runs/Feb08_19-02-05_ruche-gpu12.cluster/events.out.tfevents.1707415329.ruche-gpu12.cluster.7591.0
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:62b91494a802ce7717f6a0393156ca559317a6542f2ff37dadf904ca50ca7ae7
|
| 3 |
+
size 84095
|