Upload folder using huggingface_hub
Browse files- 23_128_e3_3e-5/README.md +63 -0
- 23_128_e3_3e-5/adapter_config.json +39 -0
- 23_128_e3_3e-5/adapter_model.safetensors +3 -0
- 23_128_e3_3e-5/all_results.json +9 -0
- 23_128_e3_3e-5/config.json +32 -0
- 23_128_e3_3e-5/merges.txt +0 -0
- 23_128_e3_3e-5/special_tokens_map.json +45 -0
- 23_128_e3_3e-5/tokenizer.json +0 -0
- 23_128_e3_3e-5/tokenizer_config.json +188 -0
- 23_128_e3_3e-5/train_results.json +9 -0
- 23_128_e3_3e-5/trainer_state.json +2556 -0
- 23_128_e3_3e-5/training_args.bin +3 -0
- 23_128_e3_3e-5/vocab.json +0 -0
23_128_e3_3e-5/README.md
ADDED
|
@@ -0,0 +1,63 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
library_name: peft
|
| 3 |
+
license: apache-2.0
|
| 4 |
+
base_model: ibm-granite/granite-3.3-8b-base
|
| 5 |
+
tags:
|
| 6 |
+
- alignment-handbook
|
| 7 |
+
- generated_from_trainer
|
| 8 |
+
datasets:
|
| 9 |
+
- data/knowledge_lora_training_data_2000
|
| 10 |
+
model-index:
|
| 11 |
+
- name: 23_128_e3_3e-5
|
| 12 |
+
results: []
|
| 13 |
+
---
|
| 14 |
+
|
| 15 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 16 |
+
should probably proofread and complete it, then remove this comment. -->
|
| 17 |
+
|
| 18 |
+
# 23_128_e3_3e-5
|
| 19 |
+
|
| 20 |
+
This model is a fine-tuned version of [ibm-granite/granite-3.3-8b-base](https://huggingface.co/ibm-granite/granite-3.3-8b-base) on the data/knowledge_lora_training_data_2000 dataset.
|
| 21 |
+
|
| 22 |
+
## Model description
|
| 23 |
+
|
| 24 |
+
More information needed
|
| 25 |
+
|
| 26 |
+
## Intended uses & limitations
|
| 27 |
+
|
| 28 |
+
More information needed
|
| 29 |
+
|
| 30 |
+
## Training and evaluation data
|
| 31 |
+
|
| 32 |
+
More information needed
|
| 33 |
+
|
| 34 |
+
## Training procedure
|
| 35 |
+
|
| 36 |
+
### Training hyperparameters
|
| 37 |
+
|
| 38 |
+
The following hyperparameters were used during training:
|
| 39 |
+
- learning_rate: 3e-05
|
| 40 |
+
- train_batch_size: 2
|
| 41 |
+
- eval_batch_size: 8
|
| 42 |
+
- seed: 42
|
| 43 |
+
- distributed_type: multi-GPU
|
| 44 |
+
- num_devices: 8
|
| 45 |
+
- gradient_accumulation_steps: 2
|
| 46 |
+
- total_train_batch_size: 32
|
| 47 |
+
- total_eval_batch_size: 64
|
| 48 |
+
- optimizer: Use adamw_torch with betas=(0.9,0.95) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
|
| 49 |
+
- lr_scheduler_type: cosine
|
| 50 |
+
- lr_scheduler_warmup_ratio: 0.05
|
| 51 |
+
- num_epochs: 3.0
|
| 52 |
+
|
| 53 |
+
### Training results
|
| 54 |
+
|
| 55 |
+
|
| 56 |
+
|
| 57 |
+
### Framework versions
|
| 58 |
+
|
| 59 |
+
- PEFT 0.15.2
|
| 60 |
+
- Transformers 4.52.4
|
| 61 |
+
- Pytorch 2.7.0+cu126
|
| 62 |
+
- Datasets 3.6.0
|
| 63 |
+
- Tokenizers 0.21.2
|
23_128_e3_3e-5/adapter_config.json
ADDED
|
@@ -0,0 +1,39 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"alpha_pattern": {},
|
| 3 |
+
"auto_mapping": null,
|
| 4 |
+
"base_model_name_or_path": "ibm-granite/granite-3.3-8b-base",
|
| 5 |
+
"bias": "none",
|
| 6 |
+
"corda_config": null,
|
| 7 |
+
"eva_config": null,
|
| 8 |
+
"exclude_modules": null,
|
| 9 |
+
"fan_in_fan_out": false,
|
| 10 |
+
"inference_mode": true,
|
| 11 |
+
"init_lora_weights": true,
|
| 12 |
+
"layer_replication": null,
|
| 13 |
+
"layers_pattern": null,
|
| 14 |
+
"layers_to_transform": null,
|
| 15 |
+
"loftq_config": {},
|
| 16 |
+
"lora_alpha": 256,
|
| 17 |
+
"lora_bias": false,
|
| 18 |
+
"lora_dropout": 0.05,
|
| 19 |
+
"megatron_config": null,
|
| 20 |
+
"megatron_core": "megatron.core",
|
| 21 |
+
"modules_to_save": null,
|
| 22 |
+
"peft_type": "LORA",
|
| 23 |
+
"r": 128,
|
| 24 |
+
"rank_pattern": {},
|
| 25 |
+
"revision": null,
|
| 26 |
+
"target_modules": [
|
| 27 |
+
"o_proj",
|
| 28 |
+
"down_proj",
|
| 29 |
+
"v_proj",
|
| 30 |
+
"q_proj",
|
| 31 |
+
"gate_proj",
|
| 32 |
+
"up_proj",
|
| 33 |
+
"k_proj"
|
| 34 |
+
],
|
| 35 |
+
"task_type": "CAUSAL_LM",
|
| 36 |
+
"trainable_token_indices": null,
|
| 37 |
+
"use_dora": false,
|
| 38 |
+
"use_rslora": false
|
| 39 |
+
}
|
23_128_e3_3e-5/adapter_model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9257a28a1ab1006de0805bb54381be30b7d4c0c8aae4a512137c526c4c5b5aad
|
| 3 |
+
size 791751704
|
23_128_e3_3e-5/all_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 3.0,
|
| 3 |
+
"total_flos": 2.607849924611539e+18,
|
| 4 |
+
"train_loss": 0.5861472732900047,
|
| 5 |
+
"train_runtime": 1168.8941,
|
| 6 |
+
"train_samples": 19155,
|
| 7 |
+
"train_samples_per_second": 49.162,
|
| 8 |
+
"train_steps_per_second": 1.537
|
| 9 |
+
}
|
23_128_e3_3e-5/config.json
ADDED
|
@@ -0,0 +1,32 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"GraniteForCausalLM"
|
| 4 |
+
],
|
| 5 |
+
"attention_bias": false,
|
| 6 |
+
"attention_dropout": 0,
|
| 7 |
+
"attention_multiplier": 0.0078125,
|
| 8 |
+
"bos_token_id": 0,
|
| 9 |
+
"embedding_multiplier": 12.0,
|
| 10 |
+
"eos_token_id": 0,
|
| 11 |
+
"hidden_act": "silu",
|
| 12 |
+
"hidden_size": 4096,
|
| 13 |
+
"initializer_range": 0.02,
|
| 14 |
+
"intermediate_size": 12800,
|
| 15 |
+
"logits_scaling": 16.0,
|
| 16 |
+
"max_position_embeddings": 131072,
|
| 17 |
+
"mlp_bias": false,
|
| 18 |
+
"model_type": "granite",
|
| 19 |
+
"num_attention_heads": 32,
|
| 20 |
+
"num_hidden_layers": 40,
|
| 21 |
+
"num_key_value_heads": 8,
|
| 22 |
+
"pad_token_id": 0,
|
| 23 |
+
"residual_multiplier": 0.22,
|
| 24 |
+
"rms_norm_eps": 1e-05,
|
| 25 |
+
"rope_scaling": null,
|
| 26 |
+
"rope_theta": 10000000.0,
|
| 27 |
+
"tie_word_embeddings": true,
|
| 28 |
+
"torch_dtype": "bfloat16",
|
| 29 |
+
"transformers_version": "4.52.4",
|
| 30 |
+
"use_cache": true,
|
| 31 |
+
"vocab_size": 49152
|
| 32 |
+
}
|
23_128_e3_3e-5/merges.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
23_128_e3_3e-5/special_tokens_map.json
ADDED
|
@@ -0,0 +1,45 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"additional_special_tokens": [
|
| 3 |
+
"<|endoftext|>",
|
| 4 |
+
"<fim_prefix>",
|
| 5 |
+
"<fim_middle>",
|
| 6 |
+
"<fim_suffix>",
|
| 7 |
+
"<fim_pad>",
|
| 8 |
+
"<filename>",
|
| 9 |
+
"<gh_stars>",
|
| 10 |
+
"<issue_start>",
|
| 11 |
+
"<issue_comment>",
|
| 12 |
+
"<issue_closed>",
|
| 13 |
+
"<jupyter_start>",
|
| 14 |
+
"<jupyter_text>",
|
| 15 |
+
"<jupyter_code>",
|
| 16 |
+
"<jupyter_output>",
|
| 17 |
+
"<empty_output>",
|
| 18 |
+
"<commit_before>",
|
| 19 |
+
"<commit_msg>",
|
| 20 |
+
"<commit_after>",
|
| 21 |
+
"<reponame>"
|
| 22 |
+
],
|
| 23 |
+
"bos_token": {
|
| 24 |
+
"content": "<|endoftext|>",
|
| 25 |
+
"lstrip": false,
|
| 26 |
+
"normalized": false,
|
| 27 |
+
"rstrip": false,
|
| 28 |
+
"single_word": false
|
| 29 |
+
},
|
| 30 |
+
"eos_token": {
|
| 31 |
+
"content": "<|endoftext|>",
|
| 32 |
+
"lstrip": false,
|
| 33 |
+
"normalized": false,
|
| 34 |
+
"rstrip": false,
|
| 35 |
+
"single_word": false
|
| 36 |
+
},
|
| 37 |
+
"pad_token": "<reponame>",
|
| 38 |
+
"unk_token": {
|
| 39 |
+
"content": "<|endoftext|>",
|
| 40 |
+
"lstrip": false,
|
| 41 |
+
"normalized": false,
|
| 42 |
+
"rstrip": false,
|
| 43 |
+
"single_word": false
|
| 44 |
+
}
|
| 45 |
+
}
|
23_128_e3_3e-5/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
23_128_e3_3e-5/tokenizer_config.json
ADDED
|
@@ -0,0 +1,188 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_prefix_space": false,
|
| 3 |
+
"added_tokens_decoder": {
|
| 4 |
+
"0": {
|
| 5 |
+
"content": "<|endoftext|>",
|
| 6 |
+
"lstrip": false,
|
| 7 |
+
"normalized": false,
|
| 8 |
+
"rstrip": false,
|
| 9 |
+
"single_word": false,
|
| 10 |
+
"special": true
|
| 11 |
+
},
|
| 12 |
+
"1": {
|
| 13 |
+
"content": "<fim_prefix>",
|
| 14 |
+
"lstrip": false,
|
| 15 |
+
"normalized": false,
|
| 16 |
+
"rstrip": false,
|
| 17 |
+
"single_word": false,
|
| 18 |
+
"special": true
|
| 19 |
+
},
|
| 20 |
+
"2": {
|
| 21 |
+
"content": "<fim_middle>",
|
| 22 |
+
"lstrip": false,
|
| 23 |
+
"normalized": false,
|
| 24 |
+
"rstrip": false,
|
| 25 |
+
"single_word": false,
|
| 26 |
+
"special": true
|
| 27 |
+
},
|
| 28 |
+
"3": {
|
| 29 |
+
"content": "<fim_suffix>",
|
| 30 |
+
"lstrip": false,
|
| 31 |
+
"normalized": false,
|
| 32 |
+
"rstrip": false,
|
| 33 |
+
"single_word": false,
|
| 34 |
+
"special": true
|
| 35 |
+
},
|
| 36 |
+
"4": {
|
| 37 |
+
"content": "<fim_pad>",
|
| 38 |
+
"lstrip": false,
|
| 39 |
+
"normalized": false,
|
| 40 |
+
"rstrip": false,
|
| 41 |
+
"single_word": false,
|
| 42 |
+
"special": true
|
| 43 |
+
},
|
| 44 |
+
"5": {
|
| 45 |
+
"content": "<filename>",
|
| 46 |
+
"lstrip": false,
|
| 47 |
+
"normalized": false,
|
| 48 |
+
"rstrip": false,
|
| 49 |
+
"single_word": false,
|
| 50 |
+
"special": true
|
| 51 |
+
},
|
| 52 |
+
"6": {
|
| 53 |
+
"content": "<gh_stars>",
|
| 54 |
+
"lstrip": false,
|
| 55 |
+
"normalized": false,
|
| 56 |
+
"rstrip": false,
|
| 57 |
+
"single_word": false,
|
| 58 |
+
"special": true
|
| 59 |
+
},
|
| 60 |
+
"7": {
|
| 61 |
+
"content": "<issue_start>",
|
| 62 |
+
"lstrip": false,
|
| 63 |
+
"normalized": false,
|
| 64 |
+
"rstrip": false,
|
| 65 |
+
"single_word": false,
|
| 66 |
+
"special": true
|
| 67 |
+
},
|
| 68 |
+
"8": {
|
| 69 |
+
"content": "<issue_comment>",
|
| 70 |
+
"lstrip": false,
|
| 71 |
+
"normalized": false,
|
| 72 |
+
"rstrip": false,
|
| 73 |
+
"single_word": false,
|
| 74 |
+
"special": true
|
| 75 |
+
},
|
| 76 |
+
"9": {
|
| 77 |
+
"content": "<issue_closed>",
|
| 78 |
+
"lstrip": false,
|
| 79 |
+
"normalized": false,
|
| 80 |
+
"rstrip": false,
|
| 81 |
+
"single_word": false,
|
| 82 |
+
"special": true
|
| 83 |
+
},
|
| 84 |
+
"10": {
|
| 85 |
+
"content": "<jupyter_start>",
|
| 86 |
+
"lstrip": false,
|
| 87 |
+
"normalized": false,
|
| 88 |
+
"rstrip": false,
|
| 89 |
+
"single_word": false,
|
| 90 |
+
"special": true
|
| 91 |
+
},
|
| 92 |
+
"11": {
|
| 93 |
+
"content": "<jupyter_text>",
|
| 94 |
+
"lstrip": false,
|
| 95 |
+
"normalized": false,
|
| 96 |
+
"rstrip": false,
|
| 97 |
+
"single_word": false,
|
| 98 |
+
"special": true
|
| 99 |
+
},
|
| 100 |
+
"12": {
|
| 101 |
+
"content": "<jupyter_code>",
|
| 102 |
+
"lstrip": false,
|
| 103 |
+
"normalized": false,
|
| 104 |
+
"rstrip": false,
|
| 105 |
+
"single_word": false,
|
| 106 |
+
"special": true
|
| 107 |
+
},
|
| 108 |
+
"13": {
|
| 109 |
+
"content": "<jupyter_output>",
|
| 110 |
+
"lstrip": false,
|
| 111 |
+
"normalized": false,
|
| 112 |
+
"rstrip": false,
|
| 113 |
+
"single_word": false,
|
| 114 |
+
"special": true
|
| 115 |
+
},
|
| 116 |
+
"14": {
|
| 117 |
+
"content": "<empty_output>",
|
| 118 |
+
"lstrip": false,
|
| 119 |
+
"normalized": false,
|
| 120 |
+
"rstrip": false,
|
| 121 |
+
"single_word": false,
|
| 122 |
+
"special": true
|
| 123 |
+
},
|
| 124 |
+
"15": {
|
| 125 |
+
"content": "<commit_before>",
|
| 126 |
+
"lstrip": false,
|
| 127 |
+
"normalized": false,
|
| 128 |
+
"rstrip": false,
|
| 129 |
+
"single_word": false,
|
| 130 |
+
"special": true
|
| 131 |
+
},
|
| 132 |
+
"16": {
|
| 133 |
+
"content": "<commit_msg>",
|
| 134 |
+
"lstrip": false,
|
| 135 |
+
"normalized": false,
|
| 136 |
+
"rstrip": false,
|
| 137 |
+
"single_word": false,
|
| 138 |
+
"special": true
|
| 139 |
+
},
|
| 140 |
+
"17": {
|
| 141 |
+
"content": "<commit_after>",
|
| 142 |
+
"lstrip": false,
|
| 143 |
+
"normalized": false,
|
| 144 |
+
"rstrip": false,
|
| 145 |
+
"single_word": false,
|
| 146 |
+
"special": true
|
| 147 |
+
},
|
| 148 |
+
"18": {
|
| 149 |
+
"content": "<reponame>",
|
| 150 |
+
"lstrip": false,
|
| 151 |
+
"normalized": false,
|
| 152 |
+
"rstrip": false,
|
| 153 |
+
"single_word": false,
|
| 154 |
+
"special": true
|
| 155 |
+
}
|
| 156 |
+
},
|
| 157 |
+
"additional_special_tokens": [
|
| 158 |
+
"<|endoftext|>",
|
| 159 |
+
"<fim_prefix>",
|
| 160 |
+
"<fim_middle>",
|
| 161 |
+
"<fim_suffix>",
|
| 162 |
+
"<fim_pad>",
|
| 163 |
+
"<filename>",
|
| 164 |
+
"<gh_stars>",
|
| 165 |
+
"<issue_start>",
|
| 166 |
+
"<issue_comment>",
|
| 167 |
+
"<issue_closed>",
|
| 168 |
+
"<jupyter_start>",
|
| 169 |
+
"<jupyter_text>",
|
| 170 |
+
"<jupyter_code>",
|
| 171 |
+
"<jupyter_output>",
|
| 172 |
+
"<empty_output>",
|
| 173 |
+
"<commit_before>",
|
| 174 |
+
"<commit_msg>",
|
| 175 |
+
"<commit_after>",
|
| 176 |
+
"<reponame>"
|
| 177 |
+
],
|
| 178 |
+
"bos_token": "<|endoftext|>",
|
| 179 |
+
"clean_up_tokenization_spaces": true,
|
| 180 |
+
"eos_token": "<|endoftext|>",
|
| 181 |
+
"extra_special_tokens": {},
|
| 182 |
+
"model_max_length": 8192,
|
| 183 |
+
"pad_token": "<reponame>",
|
| 184 |
+
"padding_side": "left",
|
| 185 |
+
"tokenizer_class": "GPT2Tokenizer",
|
| 186 |
+
"unk_token": "<|endoftext|>",
|
| 187 |
+
"vocab_size": 49152
|
| 188 |
+
}
|
23_128_e3_3e-5/train_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 3.0,
|
| 3 |
+
"total_flos": 2.607849924611539e+18,
|
| 4 |
+
"train_loss": 0.5861472732900047,
|
| 5 |
+
"train_runtime": 1168.8941,
|
| 6 |
+
"train_samples": 19155,
|
| 7 |
+
"train_samples_per_second": 49.162,
|
| 8 |
+
"train_steps_per_second": 1.537
|
| 9 |
+
}
|
23_128_e3_3e-5/trainer_state.json
ADDED
|
@@ -0,0 +1,2556 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": null,
|
| 3 |
+
"best_metric": null,
|
| 4 |
+
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 3.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 1797,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.008347245409015025,
|
| 14 |
+
"grad_norm": 0.9040147066116333,
|
| 15 |
+
"learning_rate": 1.3333333333333334e-06,
|
| 16 |
+
"loss": 1.3895,
|
| 17 |
+
"step": 5
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.01669449081803005,
|
| 21 |
+
"grad_norm": 0.909934937953949,
|
| 22 |
+
"learning_rate": 3e-06,
|
| 23 |
+
"loss": 1.3717,
|
| 24 |
+
"step": 10
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.025041736227045076,
|
| 28 |
+
"grad_norm": 0.6758604645729065,
|
| 29 |
+
"learning_rate": 4.666666666666667e-06,
|
| 30 |
+
"loss": 1.3569,
|
| 31 |
+
"step": 15
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.0333889816360601,
|
| 35 |
+
"grad_norm": 0.6147056818008423,
|
| 36 |
+
"learning_rate": 6.333333333333333e-06,
|
| 37 |
+
"loss": 1.3182,
|
| 38 |
+
"step": 20
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.041736227045075125,
|
| 42 |
+
"grad_norm": 0.4719122648239136,
|
| 43 |
+
"learning_rate": 8e-06,
|
| 44 |
+
"loss": 1.2524,
|
| 45 |
+
"step": 25
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.05008347245409015,
|
| 49 |
+
"grad_norm": 0.4834645986557007,
|
| 50 |
+
"learning_rate": 9.666666666666667e-06,
|
| 51 |
+
"loss": 1.3275,
|
| 52 |
+
"step": 30
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.05843071786310518,
|
| 56 |
+
"grad_norm": 0.528109073638916,
|
| 57 |
+
"learning_rate": 1.1333333333333334e-05,
|
| 58 |
+
"loss": 1.2624,
|
| 59 |
+
"step": 35
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 0.0667779632721202,
|
| 63 |
+
"grad_norm": 0.5277000069618225,
|
| 64 |
+
"learning_rate": 1.3000000000000001e-05,
|
| 65 |
+
"loss": 1.238,
|
| 66 |
+
"step": 40
|
| 67 |
+
},
|
| 68 |
+
{
|
| 69 |
+
"epoch": 0.07512520868113523,
|
| 70 |
+
"grad_norm": 0.4437154233455658,
|
| 71 |
+
"learning_rate": 1.4666666666666666e-05,
|
| 72 |
+
"loss": 1.2187,
|
| 73 |
+
"step": 45
|
| 74 |
+
},
|
| 75 |
+
{
|
| 76 |
+
"epoch": 0.08347245409015025,
|
| 77 |
+
"grad_norm": 0.3858715295791626,
|
| 78 |
+
"learning_rate": 1.633333333333333e-05,
|
| 79 |
+
"loss": 1.272,
|
| 80 |
+
"step": 50
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.09181969949916527,
|
| 84 |
+
"grad_norm": 0.4028352200984955,
|
| 85 |
+
"learning_rate": 1.8e-05,
|
| 86 |
+
"loss": 1.2299,
|
| 87 |
+
"step": 55
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"epoch": 0.1001669449081803,
|
| 91 |
+
"grad_norm": 0.47062796354293823,
|
| 92 |
+
"learning_rate": 1.9666666666666666e-05,
|
| 93 |
+
"loss": 1.2468,
|
| 94 |
+
"step": 60
|
| 95 |
+
},
|
| 96 |
+
{
|
| 97 |
+
"epoch": 0.10851419031719532,
|
| 98 |
+
"grad_norm": 0.3735462427139282,
|
| 99 |
+
"learning_rate": 2.1333333333333335e-05,
|
| 100 |
+
"loss": 1.2081,
|
| 101 |
+
"step": 65
|
| 102 |
+
},
|
| 103 |
+
{
|
| 104 |
+
"epoch": 0.11686143572621036,
|
| 105 |
+
"grad_norm": 0.42394939064979553,
|
| 106 |
+
"learning_rate": 2.3000000000000003e-05,
|
| 107 |
+
"loss": 1.2262,
|
| 108 |
+
"step": 70
|
| 109 |
+
},
|
| 110 |
+
{
|
| 111 |
+
"epoch": 0.12520868113522537,
|
| 112 |
+
"grad_norm": 0.4555569291114807,
|
| 113 |
+
"learning_rate": 2.4666666666666665e-05,
|
| 114 |
+
"loss": 1.2784,
|
| 115 |
+
"step": 75
|
| 116 |
+
},
|
| 117 |
+
{
|
| 118 |
+
"epoch": 0.1335559265442404,
|
| 119 |
+
"grad_norm": 0.386436402797699,
|
| 120 |
+
"learning_rate": 2.6333333333333334e-05,
|
| 121 |
+
"loss": 1.2225,
|
| 122 |
+
"step": 80
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.1419031719532554,
|
| 126 |
+
"grad_norm": 0.5518444180488586,
|
| 127 |
+
"learning_rate": 2.8e-05,
|
| 128 |
+
"loss": 1.2084,
|
| 129 |
+
"step": 85
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.15025041736227046,
|
| 133 |
+
"grad_norm": 0.4673527479171753,
|
| 134 |
+
"learning_rate": 2.966666666666667e-05,
|
| 135 |
+
"loss": 1.1771,
|
| 136 |
+
"step": 90
|
| 137 |
+
},
|
| 138 |
+
{
|
| 139 |
+
"epoch": 0.15859766277128548,
|
| 140 |
+
"grad_norm": 0.5485560297966003,
|
| 141 |
+
"learning_rate": 2.9999593545451607e-05,
|
| 142 |
+
"loss": 1.1745,
|
| 143 |
+
"step": 95
|
| 144 |
+
},
|
| 145 |
+
{
|
| 146 |
+
"epoch": 0.1669449081803005,
|
| 147 |
+
"grad_norm": 0.46298056840896606,
|
| 148 |
+
"learning_rate": 2.9997942361600667e-05,
|
| 149 |
+
"loss": 1.088,
|
| 150 |
+
"step": 100
|
| 151 |
+
},
|
| 152 |
+
{
|
| 153 |
+
"epoch": 0.17529215358931552,
|
| 154 |
+
"grad_norm": 0.516417920589447,
|
| 155 |
+
"learning_rate": 2.9995021184748996e-05,
|
| 156 |
+
"loss": 1.2131,
|
| 157 |
+
"step": 105
|
| 158 |
+
},
|
| 159 |
+
{
|
| 160 |
+
"epoch": 0.18363939899833054,
|
| 161 |
+
"grad_norm": 0.5785250067710876,
|
| 162 |
+
"learning_rate": 2.9990830262255468e-05,
|
| 163 |
+
"loss": 1.1168,
|
| 164 |
+
"step": 110
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.19198664440734559,
|
| 168 |
+
"grad_norm": 0.46146246790885925,
|
| 169 |
+
"learning_rate": 2.9985369948998253e-05,
|
| 170 |
+
"loss": 1.1687,
|
| 171 |
+
"step": 115
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 0.2003338898163606,
|
| 175 |
+
"grad_norm": 0.4953807592391968,
|
| 176 |
+
"learning_rate": 2.997864070734474e-05,
|
| 177 |
+
"loss": 1.146,
|
| 178 |
+
"step": 120
|
| 179 |
+
},
|
| 180 |
+
{
|
| 181 |
+
"epoch": 0.20868113522537562,
|
| 182 |
+
"grad_norm": 0.6573898792266846,
|
| 183 |
+
"learning_rate": 2.997064310711242e-05,
|
| 184 |
+
"loss": 1.1464,
|
| 185 |
+
"step": 125
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 0.21702838063439064,
|
| 189 |
+
"grad_norm": 0.5272989273071289,
|
| 190 |
+
"learning_rate": 2.9961377825520612e-05,
|
| 191 |
+
"loss": 1.1227,
|
| 192 |
+
"step": 130
|
| 193 |
+
},
|
| 194 |
+
{
|
| 195 |
+
"epoch": 0.22537562604340566,
|
| 196 |
+
"grad_norm": 0.5319401621818542,
|
| 197 |
+
"learning_rate": 2.9950845647133125e-05,
|
| 198 |
+
"loss": 1.1426,
|
| 199 |
+
"step": 135
|
| 200 |
+
},
|
| 201 |
+
{
|
| 202 |
+
"epoch": 0.2337228714524207,
|
| 203 |
+
"grad_norm": 0.587548553943634,
|
| 204 |
+
"learning_rate": 2.9939047463791817e-05,
|
| 205 |
+
"loss": 1.1552,
|
| 206 |
+
"step": 140
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.24207011686143573,
|
| 210 |
+
"grad_norm": 0.5197307467460632,
|
| 211 |
+
"learning_rate": 2.992598427454109e-05,
|
| 212 |
+
"loss": 1.0866,
|
| 213 |
+
"step": 145
|
| 214 |
+
},
|
| 215 |
+
{
|
| 216 |
+
"epoch": 0.25041736227045075,
|
| 217 |
+
"grad_norm": 0.6146838068962097,
|
| 218 |
+
"learning_rate": 2.9911657185543264e-05,
|
| 219 |
+
"loss": 1.1209,
|
| 220 |
+
"step": 150
|
| 221 |
+
},
|
| 222 |
+
{
|
| 223 |
+
"epoch": 0.2587646076794658,
|
| 224 |
+
"grad_norm": 0.5579906702041626,
|
| 225 |
+
"learning_rate": 2.989606740998496e-05,
|
| 226 |
+
"loss": 1.0953,
|
| 227 |
+
"step": 155
|
| 228 |
+
},
|
| 229 |
+
{
|
| 230 |
+
"epoch": 0.2671118530884808,
|
| 231 |
+
"grad_norm": 0.5035932660102844,
|
| 232 |
+
"learning_rate": 2.9879216267974306e-05,
|
| 233 |
+
"loss": 1.0758,
|
| 234 |
+
"step": 160
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"epoch": 0.27545909849749584,
|
| 238 |
+
"grad_norm": 0.7449677586555481,
|
| 239 |
+
"learning_rate": 2.986110518642921e-05,
|
| 240 |
+
"loss": 1.0847,
|
| 241 |
+
"step": 165
|
| 242 |
+
},
|
| 243 |
+
{
|
| 244 |
+
"epoch": 0.2838063439065108,
|
| 245 |
+
"grad_norm": 0.5657216906547546,
|
| 246 |
+
"learning_rate": 2.98417356989565e-05,
|
| 247 |
+
"loss": 1.0821,
|
| 248 |
+
"step": 170
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 0.2921535893155259,
|
| 252 |
+
"grad_norm": 0.5445929169654846,
|
| 253 |
+
"learning_rate": 2.9821109445722062e-05,
|
| 254 |
+
"loss": 1.0372,
|
| 255 |
+
"step": 175
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 0.3005008347245409,
|
| 259 |
+
"grad_norm": 0.6295134425163269,
|
| 260 |
+
"learning_rate": 2.9799228173311975e-05,
|
| 261 |
+
"loss": 1.0228,
|
| 262 |
+
"step": 180
|
| 263 |
+
},
|
| 264 |
+
{
|
| 265 |
+
"epoch": 0.3088480801335559,
|
| 266 |
+
"grad_norm": 0.5463770627975464,
|
| 267 |
+
"learning_rate": 2.9776093734584585e-05,
|
| 268 |
+
"loss": 1.0625,
|
| 269 |
+
"step": 185
|
| 270 |
+
},
|
| 271 |
+
{
|
| 272 |
+
"epoch": 0.31719532554257096,
|
| 273 |
+
"grad_norm": 0.5830230116844177,
|
| 274 |
+
"learning_rate": 2.9751708088513642e-05,
|
| 275 |
+
"loss": 1.0631,
|
| 276 |
+
"step": 190
|
| 277 |
+
},
|
| 278 |
+
{
|
| 279 |
+
"epoch": 0.32554257095158595,
|
| 280 |
+
"grad_norm": 0.5724889636039734,
|
| 281 |
+
"learning_rate": 2.9726073300022386e-05,
|
| 282 |
+
"loss": 1.0226,
|
| 283 |
+
"step": 195
|
| 284 |
+
},
|
| 285 |
+
{
|
| 286 |
+
"epoch": 0.333889816360601,
|
| 287 |
+
"grad_norm": 0.5786441564559937,
|
| 288 |
+
"learning_rate": 2.969919153980872e-05,
|
| 289 |
+
"loss": 1.0588,
|
| 290 |
+
"step": 200
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 0.34223706176961605,
|
| 294 |
+
"grad_norm": 0.5639914870262146,
|
| 295 |
+
"learning_rate": 2.967106508416138e-05,
|
| 296 |
+
"loss": 1.0613,
|
| 297 |
+
"step": 205
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 0.35058430717863104,
|
| 301 |
+
"grad_norm": 0.6605404019355774,
|
| 302 |
+
"learning_rate": 2.9641696314767198e-05,
|
| 303 |
+
"loss": 1.124,
|
| 304 |
+
"step": 210
|
| 305 |
+
},
|
| 306 |
+
{
|
| 307 |
+
"epoch": 0.3589315525876461,
|
| 308 |
+
"grad_norm": 0.7098401784896851,
|
| 309 |
+
"learning_rate": 2.961108771850942e-05,
|
| 310 |
+
"loss": 1.0269,
|
| 311 |
+
"step": 215
|
| 312 |
+
},
|
| 313 |
+
{
|
| 314 |
+
"epoch": 0.3672787979966611,
|
| 315 |
+
"grad_norm": 0.6752082109451294,
|
| 316 |
+
"learning_rate": 2.9579241887257117e-05,
|
| 317 |
+
"loss": 1.0193,
|
| 318 |
+
"step": 220
|
| 319 |
+
},
|
| 320 |
+
{
|
| 321 |
+
"epoch": 0.3756260434056761,
|
| 322 |
+
"grad_norm": 0.5603159070014954,
|
| 323 |
+
"learning_rate": 2.954616151764573e-05,
|
| 324 |
+
"loss": 1.0316,
|
| 325 |
+
"step": 225
|
| 326 |
+
},
|
| 327 |
+
{
|
| 328 |
+
"epoch": 0.38397328881469117,
|
| 329 |
+
"grad_norm": 0.667381227016449,
|
| 330 |
+
"learning_rate": 2.9511849410848698e-05,
|
| 331 |
+
"loss": 1.0279,
|
| 332 |
+
"step": 230
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 0.39232053422370616,
|
| 336 |
+
"grad_norm": 0.6742578744888306,
|
| 337 |
+
"learning_rate": 2.9476308472340282e-05,
|
| 338 |
+
"loss": 0.9802,
|
| 339 |
+
"step": 235
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 0.4006677796327212,
|
| 343 |
+
"grad_norm": 0.6237293481826782,
|
| 344 |
+
"learning_rate": 2.9439541711649527e-05,
|
| 345 |
+
"loss": 0.9802,
|
| 346 |
+
"step": 240
|
| 347 |
+
},
|
| 348 |
+
{
|
| 349 |
+
"epoch": 0.4090150250417362,
|
| 350 |
+
"grad_norm": 0.6732901930809021,
|
| 351 |
+
"learning_rate": 2.9401552242105423e-05,
|
| 352 |
+
"loss": 1.0088,
|
| 353 |
+
"step": 245
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"epoch": 0.41736227045075125,
|
| 357 |
+
"grad_norm": 0.6434968709945679,
|
| 358 |
+
"learning_rate": 2.9362343280573273e-05,
|
| 359 |
+
"loss": 0.9879,
|
| 360 |
+
"step": 250
|
| 361 |
+
},
|
| 362 |
+
{
|
| 363 |
+
"epoch": 0.4257095158597663,
|
| 364 |
+
"grad_norm": 0.8852955102920532,
|
| 365 |
+
"learning_rate": 2.93219181471823e-05,
|
| 366 |
+
"loss": 0.9982,
|
| 367 |
+
"step": 255
|
| 368 |
+
},
|
| 369 |
+
{
|
| 370 |
+
"epoch": 0.4340567612687813,
|
| 371 |
+
"grad_norm": 0.7065576910972595,
|
| 372 |
+
"learning_rate": 2.92802802650445e-05,
|
| 373 |
+
"loss": 0.9718,
|
| 374 |
+
"step": 260
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 0.44240400667779634,
|
| 378 |
+
"grad_norm": 0.7375466823577881,
|
| 379 |
+
"learning_rate": 2.9237433159964787e-05,
|
| 380 |
+
"loss": 0.9714,
|
| 381 |
+
"step": 265
|
| 382 |
+
},
|
| 383 |
+
{
|
| 384 |
+
"epoch": 0.4507512520868113,
|
| 385 |
+
"grad_norm": 0.8178597688674927,
|
| 386 |
+
"learning_rate": 2.9193380460142445e-05,
|
| 387 |
+
"loss": 0.9397,
|
| 388 |
+
"step": 270
|
| 389 |
+
},
|
| 390 |
+
{
|
| 391 |
+
"epoch": 0.4590984974958264,
|
| 392 |
+
"grad_norm": 0.7639552354812622,
|
| 393 |
+
"learning_rate": 2.9148125895863863e-05,
|
| 394 |
+
"loss": 0.9528,
|
| 395 |
+
"step": 275
|
| 396 |
+
},
|
| 397 |
+
{
|
| 398 |
+
"epoch": 0.4674457429048414,
|
| 399 |
+
"grad_norm": 0.7310404777526855,
|
| 400 |
+
"learning_rate": 2.9101673299186696e-05,
|
| 401 |
+
"loss": 0.9356,
|
| 402 |
+
"step": 280
|
| 403 |
+
},
|
| 404 |
+
{
|
| 405 |
+
"epoch": 0.4757929883138564,
|
| 406 |
+
"grad_norm": 0.7645027041435242,
|
| 407 |
+
"learning_rate": 2.905402660361538e-05,
|
| 408 |
+
"loss": 0.9737,
|
| 409 |
+
"step": 285
|
| 410 |
+
},
|
| 411 |
+
{
|
| 412 |
+
"epoch": 0.48414023372287146,
|
| 413 |
+
"grad_norm": 0.7474592924118042,
|
| 414 |
+
"learning_rate": 2.9005189843768023e-05,
|
| 415 |
+
"loss": 0.9323,
|
| 416 |
+
"step": 290
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 0.49248747913188645,
|
| 420 |
+
"grad_norm": 0.673616349697113,
|
| 421 |
+
"learning_rate": 2.8955167155034775e-05,
|
| 422 |
+
"loss": 0.9217,
|
| 423 |
+
"step": 295
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 0.5008347245409015,
|
| 427 |
+
"grad_norm": 0.7713244557380676,
|
| 428 |
+
"learning_rate": 2.8903962773227662e-05,
|
| 429 |
+
"loss": 0.88,
|
| 430 |
+
"step": 300
|
| 431 |
+
},
|
| 432 |
+
{
|
| 433 |
+
"epoch": 0.5091819699499165,
|
| 434 |
+
"grad_norm": 0.7548984885215759,
|
| 435 |
+
"learning_rate": 2.8851581034221898e-05,
|
| 436 |
+
"loss": 0.9675,
|
| 437 |
+
"step": 305
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 0.5175292153589316,
|
| 441 |
+
"grad_norm": 0.854121744632721,
|
| 442 |
+
"learning_rate": 2.8798026373588725e-05,
|
| 443 |
+
"loss": 0.9104,
|
| 444 |
+
"step": 310
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 0.5258764607679466,
|
| 448 |
+
"grad_norm": 0.8208110928535461,
|
| 449 |
+
"learning_rate": 2.8743303326219833e-05,
|
| 450 |
+
"loss": 0.8939,
|
| 451 |
+
"step": 315
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 0.5342237061769616,
|
| 455 |
+
"grad_norm": 0.7310402989387512,
|
| 456 |
+
"learning_rate": 2.868741652594335e-05,
|
| 457 |
+
"loss": 0.9519,
|
| 458 |
+
"step": 320
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 0.5425709515859767,
|
| 462 |
+
"grad_norm": 0.7999498844146729,
|
| 463 |
+
"learning_rate": 2.863037070513145e-05,
|
| 464 |
+
"loss": 0.9473,
|
| 465 |
+
"step": 325
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 0.5509181969949917,
|
| 469 |
+
"grad_norm": 0.7352285981178284,
|
| 470 |
+
"learning_rate": 2.8572170694299644e-05,
|
| 471 |
+
"loss": 0.8868,
|
| 472 |
+
"step": 330
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 0.5592654424040067,
|
| 476 |
+
"grad_norm": 0.7781071066856384,
|
| 477 |
+
"learning_rate": 2.8512821421697728e-05,
|
| 478 |
+
"loss": 0.8655,
|
| 479 |
+
"step": 335
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 0.5676126878130217,
|
| 483 |
+
"grad_norm": 0.7872632145881653,
|
| 484 |
+
"learning_rate": 2.845232791289248e-05,
|
| 485 |
+
"loss": 0.8928,
|
| 486 |
+
"step": 340
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 0.5759599332220368,
|
| 490 |
+
"grad_norm": 0.827265739440918,
|
| 491 |
+
"learning_rate": 2.8390695290342087e-05,
|
| 492 |
+
"loss": 0.9094,
|
| 493 |
+
"step": 345
|
| 494 |
+
},
|
| 495 |
+
{
|
| 496 |
+
"epoch": 0.5843071786310517,
|
| 497 |
+
"grad_norm": 0.7859421372413635,
|
| 498 |
+
"learning_rate": 2.8327928772962417e-05,
|
| 499 |
+
"loss": 0.8664,
|
| 500 |
+
"step": 350
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 0.5926544240400667,
|
| 504 |
+
"grad_norm": 0.9057596921920776,
|
| 505 |
+
"learning_rate": 2.8264033675685063e-05,
|
| 506 |
+
"loss": 0.9337,
|
| 507 |
+
"step": 355
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 0.6010016694490818,
|
| 511 |
+
"grad_norm": 0.8775967955589294,
|
| 512 |
+
"learning_rate": 2.8199015409007304e-05,
|
| 513 |
+
"loss": 0.9012,
|
| 514 |
+
"step": 360
|
| 515 |
+
},
|
| 516 |
+
{
|
| 517 |
+
"epoch": 0.6093489148580968,
|
| 518 |
+
"grad_norm": 0.8351387977600098,
|
| 519 |
+
"learning_rate": 2.813287947853395e-05,
|
| 520 |
+
"loss": 0.8897,
|
| 521 |
+
"step": 365
|
| 522 |
+
},
|
| 523 |
+
{
|
| 524 |
+
"epoch": 0.6176961602671118,
|
| 525 |
+
"grad_norm": 0.8018450140953064,
|
| 526 |
+
"learning_rate": 2.8065631484511136e-05,
|
| 527 |
+
"loss": 0.8,
|
| 528 |
+
"step": 370
|
| 529 |
+
},
|
| 530 |
+
{
|
| 531 |
+
"epoch": 0.6260434056761269,
|
| 532 |
+
"grad_norm": 0.7724934816360474,
|
| 533 |
+
"learning_rate": 2.799727712135212e-05,
|
| 534 |
+
"loss": 0.8606,
|
| 535 |
+
"step": 375
|
| 536 |
+
},
|
| 537 |
+
{
|
| 538 |
+
"epoch": 0.6343906510851419,
|
| 539 |
+
"grad_norm": 0.8219337463378906,
|
| 540 |
+
"learning_rate": 2.7927822177155075e-05,
|
| 541 |
+
"loss": 0.8068,
|
| 542 |
+
"step": 380
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 0.6427378964941569,
|
| 546 |
+
"grad_norm": 0.8023366332054138,
|
| 547 |
+
"learning_rate": 2.785727253321296e-05,
|
| 548 |
+
"loss": 0.873,
|
| 549 |
+
"step": 385
|
| 550 |
+
},
|
| 551 |
+
{
|
| 552 |
+
"epoch": 0.6510851419031719,
|
| 553 |
+
"grad_norm": 0.8897063136100769,
|
| 554 |
+
"learning_rate": 2.7785634163515543e-05,
|
| 555 |
+
"loss": 0.7923,
|
| 556 |
+
"step": 390
|
| 557 |
+
},
|
| 558 |
+
{
|
| 559 |
+
"epoch": 0.659432387312187,
|
| 560 |
+
"grad_norm": 0.8423841595649719,
|
| 561 |
+
"learning_rate": 2.7712913134243493e-05,
|
| 562 |
+
"loss": 0.8802,
|
| 563 |
+
"step": 395
|
| 564 |
+
},
|
| 565 |
+
{
|
| 566 |
+
"epoch": 0.667779632721202,
|
| 567 |
+
"grad_norm": 0.8655575513839722,
|
| 568 |
+
"learning_rate": 2.763911560325473e-05,
|
| 569 |
+
"loss": 0.8314,
|
| 570 |
+
"step": 400
|
| 571 |
+
},
|
| 572 |
+
{
|
| 573 |
+
"epoch": 0.676126878130217,
|
| 574 |
+
"grad_norm": 0.8509688377380371,
|
| 575 |
+
"learning_rate": 2.7564247819562993e-05,
|
| 576 |
+
"loss": 0.7872,
|
| 577 |
+
"step": 405
|
| 578 |
+
},
|
| 579 |
+
{
|
| 580 |
+
"epoch": 0.6844741235392321,
|
| 581 |
+
"grad_norm": 0.8914806842803955,
|
| 582 |
+
"learning_rate": 2.748831612280868e-05,
|
| 583 |
+
"loss": 0.8268,
|
| 584 |
+
"step": 410
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 0.6928213689482471,
|
| 588 |
+
"grad_norm": 0.9779593348503113,
|
| 589 |
+
"learning_rate": 2.7411326942722023e-05,
|
| 590 |
+
"loss": 0.8363,
|
| 591 |
+
"step": 415
|
| 592 |
+
},
|
| 593 |
+
{
|
| 594 |
+
"epoch": 0.7011686143572621,
|
| 595 |
+
"grad_norm": 0.8568240404129028,
|
| 596 |
+
"learning_rate": 2.733328679857862e-05,
|
| 597 |
+
"loss": 0.8115,
|
| 598 |
+
"step": 420
|
| 599 |
+
},
|
| 600 |
+
{
|
| 601 |
+
"epoch": 0.7095158597662772,
|
| 602 |
+
"grad_norm": 0.9036301374435425,
|
| 603 |
+
"learning_rate": 2.725420229864744e-05,
|
| 604 |
+
"loss": 0.7993,
|
| 605 |
+
"step": 425
|
| 606 |
+
},
|
| 607 |
+
{
|
| 608 |
+
"epoch": 0.7178631051752922,
|
| 609 |
+
"grad_norm": 0.995337188243866,
|
| 610 |
+
"learning_rate": 2.717408013963119e-05,
|
| 611 |
+
"loss": 0.8094,
|
| 612 |
+
"step": 430
|
| 613 |
+
},
|
| 614 |
+
{
|
| 615 |
+
"epoch": 0.7262103505843072,
|
| 616 |
+
"grad_norm": 0.8242852687835693,
|
| 617 |
+
"learning_rate": 2.7092927106099293e-05,
|
| 618 |
+
"loss": 0.7878,
|
| 619 |
+
"step": 435
|
| 620 |
+
},
|
| 621 |
+
{
|
| 622 |
+
"epoch": 0.7345575959933222,
|
| 623 |
+
"grad_norm": 0.8258833885192871,
|
| 624 |
+
"learning_rate": 2.701075006991338e-05,
|
| 625 |
+
"loss": 0.8256,
|
| 626 |
+
"step": 440
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 0.7429048414023373,
|
| 630 |
+
"grad_norm": 0.9491726756095886,
|
| 631 |
+
"learning_rate": 2.692755598964538e-05,
|
| 632 |
+
"loss": 0.7639,
|
| 633 |
+
"step": 445
|
| 634 |
+
},
|
| 635 |
+
{
|
| 636 |
+
"epoch": 0.7512520868113522,
|
| 637 |
+
"grad_norm": 0.8782659769058228,
|
| 638 |
+
"learning_rate": 2.68433519099883e-05,
|
| 639 |
+
"loss": 0.7953,
|
| 640 |
+
"step": 450
|
| 641 |
+
},
|
| 642 |
+
{
|
| 643 |
+
"epoch": 0.7595993322203672,
|
| 644 |
+
"grad_norm": 0.9230162501335144,
|
| 645 |
+
"learning_rate": 2.6758144961159687e-05,
|
| 646 |
+
"loss": 0.8009,
|
| 647 |
+
"step": 455
|
| 648 |
+
},
|
| 649 |
+
{
|
| 650 |
+
"epoch": 0.7679465776293823,
|
| 651 |
+
"grad_norm": 0.9994267225265503,
|
| 652 |
+
"learning_rate": 2.667194235829786e-05,
|
| 653 |
+
"loss": 0.7901,
|
| 654 |
+
"step": 460
|
| 655 |
+
},
|
| 656 |
+
{
|
| 657 |
+
"epoch": 0.7762938230383973,
|
| 658 |
+
"grad_norm": 0.8867586255073547,
|
| 659 |
+
"learning_rate": 2.6584751400850942e-05,
|
| 660 |
+
"loss": 0.8361,
|
| 661 |
+
"step": 465
|
| 662 |
+
},
|
| 663 |
+
{
|
| 664 |
+
"epoch": 0.7846410684474123,
|
| 665 |
+
"grad_norm": 0.8926465511322021,
|
| 666 |
+
"learning_rate": 2.649657947195877e-05,
|
| 667 |
+
"loss": 0.7942,
|
| 668 |
+
"step": 470
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 0.7929883138564274,
|
| 672 |
+
"grad_norm": 0.9240365028381348,
|
| 673 |
+
"learning_rate": 2.6407434037827702e-05,
|
| 674 |
+
"loss": 0.7589,
|
| 675 |
+
"step": 475
|
| 676 |
+
},
|
| 677 |
+
{
|
| 678 |
+
"epoch": 0.8013355592654424,
|
| 679 |
+
"grad_norm": 0.9446412324905396,
|
| 680 |
+
"learning_rate": 2.631732264709838e-05,
|
| 681 |
+
"loss": 0.7213,
|
| 682 |
+
"step": 480
|
| 683 |
+
},
|
| 684 |
+
{
|
| 685 |
+
"epoch": 0.8096828046744574,
|
| 686 |
+
"grad_norm": 0.8857684135437012,
|
| 687 |
+
"learning_rate": 2.6226252930206557e-05,
|
| 688 |
+
"loss": 0.7653,
|
| 689 |
+
"step": 485
|
| 690 |
+
},
|
| 691 |
+
{
|
| 692 |
+
"epoch": 0.8180300500834724,
|
| 693 |
+
"grad_norm": 0.9263355731964111,
|
| 694 |
+
"learning_rate": 2.613423259873694e-05,
|
| 695 |
+
"loss": 0.7591,
|
| 696 |
+
"step": 490
|
| 697 |
+
},
|
| 698 |
+
{
|
| 699 |
+
"epoch": 0.8263772954924875,
|
| 700 |
+
"grad_norm": 0.9910674095153809,
|
| 701 |
+
"learning_rate": 2.6041269444770214e-05,
|
| 702 |
+
"loss": 0.7392,
|
| 703 |
+
"step": 495
|
| 704 |
+
},
|
| 705 |
+
{
|
| 706 |
+
"epoch": 0.8347245409015025,
|
| 707 |
+
"grad_norm": 0.9579534530639648,
|
| 708 |
+
"learning_rate": 2.5947371340223206e-05,
|
| 709 |
+
"loss": 0.8105,
|
| 710 |
+
"step": 500
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 0.8430717863105175,
|
| 714 |
+
"grad_norm": 0.8993116021156311,
|
| 715 |
+
"learning_rate": 2.585254623618231e-05,
|
| 716 |
+
"loss": 0.7788,
|
| 717 |
+
"step": 505
|
| 718 |
+
},
|
| 719 |
+
{
|
| 720 |
+
"epoch": 0.8514190317195326,
|
| 721 |
+
"grad_norm": 0.9423431158065796,
|
| 722 |
+
"learning_rate": 2.5756802162230227e-05,
|
| 723 |
+
"loss": 0.7344,
|
| 724 |
+
"step": 510
|
| 725 |
+
},
|
| 726 |
+
{
|
| 727 |
+
"epoch": 0.8597662771285476,
|
| 728 |
+
"grad_norm": 0.9205474853515625,
|
| 729 |
+
"learning_rate": 2.566014722576601e-05,
|
| 730 |
+
"loss": 0.7816,
|
| 731 |
+
"step": 515
|
| 732 |
+
},
|
| 733 |
+
{
|
| 734 |
+
"epoch": 0.8681135225375626,
|
| 735 |
+
"grad_norm": 0.9966205954551697,
|
| 736 |
+
"learning_rate": 2.556258961131857e-05,
|
| 737 |
+
"loss": 0.7237,
|
| 738 |
+
"step": 520
|
| 739 |
+
},
|
| 740 |
+
{
|
| 741 |
+
"epoch": 0.8764607679465777,
|
| 742 |
+
"grad_norm": 0.9185479283332825,
|
| 743 |
+
"learning_rate": 2.5464137579853616e-05,
|
| 744 |
+
"loss": 0.7105,
|
| 745 |
+
"step": 525
|
| 746 |
+
},
|
| 747 |
+
{
|
| 748 |
+
"epoch": 0.8848080133555927,
|
| 749 |
+
"grad_norm": 1.0423331260681152,
|
| 750 |
+
"learning_rate": 2.5364799468074126e-05,
|
| 751 |
+
"loss": 0.75,
|
| 752 |
+
"step": 530
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 0.8931552587646077,
|
| 756 |
+
"grad_norm": 0.9681307673454285,
|
| 757 |
+
"learning_rate": 2.5264583687714434e-05,
|
| 758 |
+
"loss": 0.7251,
|
| 759 |
+
"step": 535
|
| 760 |
+
},
|
| 761 |
+
{
|
| 762 |
+
"epoch": 0.9015025041736227,
|
| 763 |
+
"grad_norm": 0.9622064828872681,
|
| 764 |
+
"learning_rate": 2.516349872482793e-05,
|
| 765 |
+
"loss": 0.7701,
|
| 766 |
+
"step": 540
|
| 767 |
+
},
|
| 768 |
+
{
|
| 769 |
+
"epoch": 0.9098497495826378,
|
| 770 |
+
"grad_norm": 0.9897941946983337,
|
| 771 |
+
"learning_rate": 2.5061553139068487e-05,
|
| 772 |
+
"loss": 0.7223,
|
| 773 |
+
"step": 545
|
| 774 |
+
},
|
| 775 |
+
{
|
| 776 |
+
"epoch": 0.9181969949916527,
|
| 777 |
+
"grad_norm": 0.9223009347915649,
|
| 778 |
+
"learning_rate": 2.495875556296563e-05,
|
| 779 |
+
"loss": 0.7251,
|
| 780 |
+
"step": 550
|
| 781 |
+
},
|
| 782 |
+
{
|
| 783 |
+
"epoch": 0.9265442404006677,
|
| 784 |
+
"grad_norm": 0.9754142761230469,
|
| 785 |
+
"learning_rate": 2.485511470119358e-05,
|
| 786 |
+
"loss": 0.6836,
|
| 787 |
+
"step": 555
|
| 788 |
+
},
|
| 789 |
+
{
|
| 790 |
+
"epoch": 0.9348914858096828,
|
| 791 |
+
"grad_norm": 0.8462454080581665,
|
| 792 |
+
"learning_rate": 2.4750639329834135e-05,
|
| 793 |
+
"loss": 0.7613,
|
| 794 |
+
"step": 560
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 0.9432387312186978,
|
| 798 |
+
"grad_norm": 0.902420163154602,
|
| 799 |
+
"learning_rate": 2.4645338295633553e-05,
|
| 800 |
+
"loss": 0.7149,
|
| 801 |
+
"step": 565
|
| 802 |
+
},
|
| 803 |
+
{
|
| 804 |
+
"epoch": 0.9515859766277128,
|
| 805 |
+
"grad_norm": 1.0617609024047852,
|
| 806 |
+
"learning_rate": 2.4539220515253406e-05,
|
| 807 |
+
"loss": 0.7,
|
| 808 |
+
"step": 570
|
| 809 |
+
},
|
| 810 |
+
{
|
| 811 |
+
"epoch": 0.9599332220367279,
|
| 812 |
+
"grad_norm": 1.099102258682251,
|
| 813 |
+
"learning_rate": 2.4432294974515553e-05,
|
| 814 |
+
"loss": 0.7311,
|
| 815 |
+
"step": 575
|
| 816 |
+
},
|
| 817 |
+
{
|
| 818 |
+
"epoch": 0.9682804674457429,
|
| 819 |
+
"grad_norm": 1.216670036315918,
|
| 820 |
+
"learning_rate": 2.4324570727641247e-05,
|
| 821 |
+
"loss": 0.7207,
|
| 822 |
+
"step": 580
|
| 823 |
+
},
|
| 824 |
+
{
|
| 825 |
+
"epoch": 0.9766277128547579,
|
| 826 |
+
"grad_norm": 0.9607879519462585,
|
| 827 |
+
"learning_rate": 2.421605689648441e-05,
|
| 828 |
+
"loss": 0.713,
|
| 829 |
+
"step": 585
|
| 830 |
+
},
|
| 831 |
+
{
|
| 832 |
+
"epoch": 0.9849749582637729,
|
| 833 |
+
"grad_norm": 1.008848786354065,
|
| 834 |
+
"learning_rate": 2.4106762669759265e-05,
|
| 835 |
+
"loss": 0.6883,
|
| 836 |
+
"step": 590
|
| 837 |
+
},
|
| 838 |
+
{
|
| 839 |
+
"epoch": 0.993322203672788,
|
| 840 |
+
"grad_norm": 0.8814200758934021,
|
| 841 |
+
"learning_rate": 2.3996697302262195e-05,
|
| 842 |
+
"loss": 0.6938,
|
| 843 |
+
"step": 595
|
| 844 |
+
},
|
| 845 |
+
{
|
| 846 |
+
"epoch": 1.001669449081803,
|
| 847 |
+
"grad_norm": 0.9263659119606018,
|
| 848 |
+
"learning_rate": 2.3885870114088117e-05,
|
| 849 |
+
"loss": 0.6666,
|
| 850 |
+
"step": 600
|
| 851 |
+
},
|
| 852 |
+
{
|
| 853 |
+
"epoch": 1.010016694490818,
|
| 854 |
+
"grad_norm": 1.1444543600082397,
|
| 855 |
+
"learning_rate": 2.3774290489841267e-05,
|
| 856 |
+
"loss": 0.6429,
|
| 857 |
+
"step": 605
|
| 858 |
+
},
|
| 859 |
+
{
|
| 860 |
+
"epoch": 1.018363939899833,
|
| 861 |
+
"grad_norm": 0.9786504507064819,
|
| 862 |
+
"learning_rate": 2.3661967877840506e-05,
|
| 863 |
+
"loss": 0.604,
|
| 864 |
+
"step": 610
|
| 865 |
+
},
|
| 866 |
+
{
|
| 867 |
+
"epoch": 1.026711185308848,
|
| 868 |
+
"grad_norm": 1.011283278465271,
|
| 869 |
+
"learning_rate": 2.354891178931929e-05,
|
| 870 |
+
"loss": 0.6637,
|
| 871 |
+
"step": 615
|
| 872 |
+
},
|
| 873 |
+
{
|
| 874 |
+
"epoch": 1.0350584307178632,
|
| 875 |
+
"grad_norm": 0.8889780640602112,
|
| 876 |
+
"learning_rate": 2.3435131797620254e-05,
|
| 877 |
+
"loss": 0.5721,
|
| 878 |
+
"step": 620
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 1.0434056761268782,
|
| 882 |
+
"grad_norm": 1.0436961650848389,
|
| 883 |
+
"learning_rate": 2.3320637537384583e-05,
|
| 884 |
+
"loss": 0.6399,
|
| 885 |
+
"step": 625
|
| 886 |
+
},
|
| 887 |
+
{
|
| 888 |
+
"epoch": 1.0517529215358932,
|
| 889 |
+
"grad_norm": 1.0458881855010986,
|
| 890 |
+
"learning_rate": 2.320543870373615e-05,
|
| 891 |
+
"loss": 0.6246,
|
| 892 |
+
"step": 630
|
| 893 |
+
},
|
| 894 |
+
{
|
| 895 |
+
"epoch": 1.0601001669449082,
|
| 896 |
+
"grad_norm": 1.055764079093933,
|
| 897 |
+
"learning_rate": 2.3089545051460568e-05,
|
| 898 |
+
"loss": 0.5927,
|
| 899 |
+
"step": 635
|
| 900 |
+
},
|
| 901 |
+
{
|
| 902 |
+
"epoch": 1.0684474123539232,
|
| 903 |
+
"grad_norm": 1.1730031967163086,
|
| 904 |
+
"learning_rate": 2.297296639417918e-05,
|
| 905 |
+
"loss": 0.5758,
|
| 906 |
+
"step": 640
|
| 907 |
+
},
|
| 908 |
+
{
|
| 909 |
+
"epoch": 1.0767946577629381,
|
| 910 |
+
"grad_norm": 1.1036263704299927,
|
| 911 |
+
"learning_rate": 2.285571260351806e-05,
|
| 912 |
+
"loss": 0.5932,
|
| 913 |
+
"step": 645
|
| 914 |
+
},
|
| 915 |
+
{
|
| 916 |
+
"epoch": 1.0851419031719534,
|
| 917 |
+
"grad_norm": 0.9903286099433899,
|
| 918 |
+
"learning_rate": 2.273779360827209e-05,
|
| 919 |
+
"loss": 0.6289,
|
| 920 |
+
"step": 650
|
| 921 |
+
},
|
| 922 |
+
{
|
| 923 |
+
"epoch": 1.0934891485809684,
|
| 924 |
+
"grad_norm": 0.9977178573608398,
|
| 925 |
+
"learning_rate": 2.2619219393564237e-05,
|
| 926 |
+
"loss": 0.5681,
|
| 927 |
+
"step": 655
|
| 928 |
+
},
|
| 929 |
+
{
|
| 930 |
+
"epoch": 1.1018363939899833,
|
| 931 |
+
"grad_norm": 1.082796573638916,
|
| 932 |
+
"learning_rate": 2.25e-05,
|
| 933 |
+
"loss": 0.612,
|
| 934 |
+
"step": 660
|
| 935 |
+
},
|
| 936 |
+
{
|
| 937 |
+
"epoch": 1.1101836393989983,
|
| 938 |
+
"grad_norm": 1.0099008083343506,
|
| 939 |
+
"learning_rate": 2.238014552281724e-05,
|
| 940 |
+
"loss": 0.5493,
|
| 941 |
+
"step": 665
|
| 942 |
+
},
|
| 943 |
+
{
|
| 944 |
+
"epoch": 1.1185308848080133,
|
| 945 |
+
"grad_norm": 0.968538224697113,
|
| 946 |
+
"learning_rate": 2.225966611103129e-05,
|
| 947 |
+
"loss": 0.5455,
|
| 948 |
+
"step": 670
|
| 949 |
+
},
|
| 950 |
+
{
|
| 951 |
+
"epoch": 1.1268781302170283,
|
| 952 |
+
"grad_norm": 1.196339726448059,
|
| 953 |
+
"learning_rate": 2.2138571966575597e-05,
|
| 954 |
+
"loss": 0.572,
|
| 955 |
+
"step": 675
|
| 956 |
+
},
|
| 957 |
+
{
|
| 958 |
+
"epoch": 1.1352253756260433,
|
| 959 |
+
"grad_norm": 0.993902325630188,
|
| 960 |
+
"learning_rate": 2.2016873343437808e-05,
|
| 961 |
+
"loss": 0.5941,
|
| 962 |
+
"step": 680
|
| 963 |
+
},
|
| 964 |
+
{
|
| 965 |
+
"epoch": 1.1435726210350585,
|
| 966 |
+
"grad_norm": 0.9588627219200134,
|
| 967 |
+
"learning_rate": 2.1894580546791532e-05,
|
| 968 |
+
"loss": 0.5795,
|
| 969 |
+
"step": 685
|
| 970 |
+
},
|
| 971 |
+
{
|
| 972 |
+
"epoch": 1.1519198664440735,
|
| 973 |
+
"grad_norm": 1.1982269287109375,
|
| 974 |
+
"learning_rate": 2.1771703932123667e-05,
|
| 975 |
+
"loss": 0.5664,
|
| 976 |
+
"step": 690
|
| 977 |
+
},
|
| 978 |
+
{
|
| 979 |
+
"epoch": 1.1602671118530885,
|
| 980 |
+
"grad_norm": 1.1193559169769287,
|
| 981 |
+
"learning_rate": 2.164825390435756e-05,
|
| 982 |
+
"loss": 0.5879,
|
| 983 |
+
"step": 695
|
| 984 |
+
},
|
| 985 |
+
{
|
| 986 |
+
"epoch": 1.1686143572621035,
|
| 987 |
+
"grad_norm": 1.071407675743103,
|
| 988 |
+
"learning_rate": 2.1524240916971928e-05,
|
| 989 |
+
"loss": 0.5641,
|
| 990 |
+
"step": 700
|
| 991 |
+
},
|
| 992 |
+
{
|
| 993 |
+
"epoch": 1.1769616026711185,
|
| 994 |
+
"grad_norm": 1.095268964767456,
|
| 995 |
+
"learning_rate": 2.1399675471115662e-05,
|
| 996 |
+
"loss": 0.5898,
|
| 997 |
+
"step": 705
|
| 998 |
+
},
|
| 999 |
+
{
|
| 1000 |
+
"epoch": 1.1853088480801335,
|
| 1001 |
+
"grad_norm": 1.2066850662231445,
|
| 1002 |
+
"learning_rate": 2.1274568114718653e-05,
|
| 1003 |
+
"loss": 0.5522,
|
| 1004 |
+
"step": 710
|
| 1005 |
+
},
|
| 1006 |
+
{
|
| 1007 |
+
"epoch": 1.1936560934891487,
|
| 1008 |
+
"grad_norm": 1.112568736076355,
|
| 1009 |
+
"learning_rate": 2.1148929441598573e-05,
|
| 1010 |
+
"loss": 0.5748,
|
| 1011 |
+
"step": 715
|
| 1012 |
+
},
|
| 1013 |
+
{
|
| 1014 |
+
"epoch": 1.2020033388981637,
|
| 1015 |
+
"grad_norm": 1.1213324069976807,
|
| 1016 |
+
"learning_rate": 2.1022770090563854e-05,
|
| 1017 |
+
"loss": 0.5146,
|
| 1018 |
+
"step": 720
|
| 1019 |
+
},
|
| 1020 |
+
{
|
| 1021 |
+
"epoch": 1.2103505843071787,
|
| 1022 |
+
"grad_norm": 1.0905853509902954,
|
| 1023 |
+
"learning_rate": 2.0896100744512773e-05,
|
| 1024 |
+
"loss": 0.5534,
|
| 1025 |
+
"step": 725
|
| 1026 |
+
},
|
| 1027 |
+
{
|
| 1028 |
+
"epoch": 1.2186978297161937,
|
| 1029 |
+
"grad_norm": 1.0488073825836182,
|
| 1030 |
+
"learning_rate": 2.076893212952889e-05,
|
| 1031 |
+
"loss": 0.5217,
|
| 1032 |
+
"step": 730
|
| 1033 |
+
},
|
| 1034 |
+
{
|
| 1035 |
+
"epoch": 1.2270450751252087,
|
| 1036 |
+
"grad_norm": 1.1673659086227417,
|
| 1037 |
+
"learning_rate": 2.0641275013972757e-05,
|
| 1038 |
+
"loss": 0.5516,
|
| 1039 |
+
"step": 735
|
| 1040 |
+
},
|
| 1041 |
+
{
|
| 1042 |
+
"epoch": 1.2353923205342237,
|
| 1043 |
+
"grad_norm": 1.0526152849197388,
|
| 1044 |
+
"learning_rate": 2.0513140207570094e-05,
|
| 1045 |
+
"loss": 0.5289,
|
| 1046 |
+
"step": 740
|
| 1047 |
+
},
|
| 1048 |
+
{
|
| 1049 |
+
"epoch": 1.2437395659432386,
|
| 1050 |
+
"grad_norm": 1.0709412097930908,
|
| 1051 |
+
"learning_rate": 2.038453856049643e-05,
|
| 1052 |
+
"loss": 0.5635,
|
| 1053 |
+
"step": 745
|
| 1054 |
+
},
|
| 1055 |
+
{
|
| 1056 |
+
"epoch": 1.2520868113522536,
|
| 1057 |
+
"grad_norm": 1.1130602359771729,
|
| 1058 |
+
"learning_rate": 2.0255480962458355e-05,
|
| 1059 |
+
"loss": 0.5708,
|
| 1060 |
+
"step": 750
|
| 1061 |
+
},
|
| 1062 |
+
{
|
| 1063 |
+
"epoch": 1.2604340567612689,
|
| 1064 |
+
"grad_norm": 1.1127012968063354,
|
| 1065 |
+
"learning_rate": 2.0125978341771377e-05,
|
| 1066 |
+
"loss": 0.5443,
|
| 1067 |
+
"step": 755
|
| 1068 |
+
},
|
| 1069 |
+
{
|
| 1070 |
+
"epoch": 1.2687813021702838,
|
| 1071 |
+
"grad_norm": 1.0472954511642456,
|
| 1072 |
+
"learning_rate": 1.9996041664434563e-05,
|
| 1073 |
+
"loss": 0.5923,
|
| 1074 |
+
"step": 760
|
| 1075 |
+
},
|
| 1076 |
+
{
|
| 1077 |
+
"epoch": 1.2771285475792988,
|
| 1078 |
+
"grad_norm": 1.062038540840149,
|
| 1079 |
+
"learning_rate": 1.9865681933201938e-05,
|
| 1080 |
+
"loss": 0.5701,
|
| 1081 |
+
"step": 765
|
| 1082 |
+
},
|
| 1083 |
+
{
|
| 1084 |
+
"epoch": 1.2854757929883138,
|
| 1085 |
+
"grad_norm": 1.0011075735092163,
|
| 1086 |
+
"learning_rate": 1.9734910186650815e-05,
|
| 1087 |
+
"loss": 0.5216,
|
| 1088 |
+
"step": 770
|
| 1089 |
+
},
|
| 1090 |
+
{
|
| 1091 |
+
"epoch": 1.2938230383973288,
|
| 1092 |
+
"grad_norm": 1.0229215621948242,
|
| 1093 |
+
"learning_rate": 1.960373749824706e-05,
|
| 1094 |
+
"loss": 0.5705,
|
| 1095 |
+
"step": 775
|
| 1096 |
+
},
|
| 1097 |
+
{
|
| 1098 |
+
"epoch": 1.302170283806344,
|
| 1099 |
+
"grad_norm": 0.9941195845603943,
|
| 1100 |
+
"learning_rate": 1.947217497540742e-05,
|
| 1101 |
+
"loss": 0.5419,
|
| 1102 |
+
"step": 780
|
| 1103 |
+
},
|
| 1104 |
+
{
|
| 1105 |
+
"epoch": 1.310517529215359,
|
| 1106 |
+
"grad_norm": 1.0428767204284668,
|
| 1107 |
+
"learning_rate": 1.934023375855897e-05,
|
| 1108 |
+
"loss": 0.5396,
|
| 1109 |
+
"step": 785
|
| 1110 |
+
},
|
| 1111 |
+
{
|
| 1112 |
+
"epoch": 1.318864774624374,
|
| 1113 |
+
"grad_norm": 1.0899183750152588,
|
| 1114 |
+
"learning_rate": 1.9207925020195767e-05,
|
| 1115 |
+
"loss": 0.5365,
|
| 1116 |
+
"step": 790
|
| 1117 |
+
},
|
| 1118 |
+
{
|
| 1119 |
+
"epoch": 1.327212020033389,
|
| 1120 |
+
"grad_norm": 1.1456999778747559,
|
| 1121 |
+
"learning_rate": 1.9075259963932786e-05,
|
| 1122 |
+
"loss": 0.5084,
|
| 1123 |
+
"step": 795
|
| 1124 |
+
},
|
| 1125 |
+
{
|
| 1126 |
+
"epoch": 1.335559265442404,
|
| 1127 |
+
"grad_norm": 1.1160317659378052,
|
| 1128 |
+
"learning_rate": 1.8942249823557226e-05,
|
| 1129 |
+
"loss": 0.5312,
|
| 1130 |
+
"step": 800
|
| 1131 |
+
},
|
| 1132 |
+
{
|
| 1133 |
+
"epoch": 1.343906510851419,
|
| 1134 |
+
"grad_norm": 1.1088972091674805,
|
| 1135 |
+
"learning_rate": 1.8808905862077258e-05,
|
| 1136 |
+
"loss": 0.5373,
|
| 1137 |
+
"step": 805
|
| 1138 |
+
},
|
| 1139 |
+
{
|
| 1140 |
+
"epoch": 1.352253756260434,
|
| 1141 |
+
"grad_norm": 1.0228931903839111,
|
| 1142 |
+
"learning_rate": 1.8675239370768283e-05,
|
| 1143 |
+
"loss": 0.5208,
|
| 1144 |
+
"step": 810
|
| 1145 |
+
},
|
| 1146 |
+
{
|
| 1147 |
+
"epoch": 1.360601001669449,
|
| 1148 |
+
"grad_norm": 1.1264790296554565,
|
| 1149 |
+
"learning_rate": 1.8541261668216836e-05,
|
| 1150 |
+
"loss": 0.5309,
|
| 1151 |
+
"step": 815
|
| 1152 |
+
},
|
| 1153 |
+
{
|
| 1154 |
+
"epoch": 1.3689482470784642,
|
| 1155 |
+
"grad_norm": 1.0168386697769165,
|
| 1156 |
+
"learning_rate": 1.8406984099362136e-05,
|
| 1157 |
+
"loss": 0.5546,
|
| 1158 |
+
"step": 820
|
| 1159 |
+
},
|
| 1160 |
+
{
|
| 1161 |
+
"epoch": 1.3772954924874792,
|
| 1162 |
+
"grad_norm": 1.1559392213821411,
|
| 1163 |
+
"learning_rate": 1.8272418034535425e-05,
|
| 1164 |
+
"loss": 0.516,
|
| 1165 |
+
"step": 825
|
| 1166 |
+
},
|
| 1167 |
+
{
|
| 1168 |
+
"epoch": 1.3856427378964942,
|
| 1169 |
+
"grad_norm": 1.1523678302764893,
|
| 1170 |
+
"learning_rate": 1.813757486849715e-05,
|
| 1171 |
+
"loss": 0.4882,
|
| 1172 |
+
"step": 830
|
| 1173 |
+
},
|
| 1174 |
+
{
|
| 1175 |
+
"epoch": 1.3939899833055092,
|
| 1176 |
+
"grad_norm": 1.0115946531295776,
|
| 1177 |
+
"learning_rate": 1.8002466019472088e-05,
|
| 1178 |
+
"loss": 0.5385,
|
| 1179 |
+
"step": 835
|
| 1180 |
+
},
|
| 1181 |
+
{
|
| 1182 |
+
"epoch": 1.4023372287145242,
|
| 1183 |
+
"grad_norm": 1.1088765859603882,
|
| 1184 |
+
"learning_rate": 1.7867102928182477e-05,
|
| 1185 |
+
"loss": 0.5429,
|
| 1186 |
+
"step": 840
|
| 1187 |
+
},
|
| 1188 |
+
{
|
| 1189 |
+
"epoch": 1.4106844741235394,
|
| 1190 |
+
"grad_norm": 1.2302361726760864,
|
| 1191 |
+
"learning_rate": 1.773149705687922e-05,
|
| 1192 |
+
"loss": 0.4745,
|
| 1193 |
+
"step": 845
|
| 1194 |
+
},
|
| 1195 |
+
{
|
| 1196 |
+
"epoch": 1.4190317195325544,
|
| 1197 |
+
"grad_norm": 1.1574212312698364,
|
| 1198 |
+
"learning_rate": 1.7595659888371312e-05,
|
| 1199 |
+
"loss": 0.4967,
|
| 1200 |
+
"step": 850
|
| 1201 |
+
},
|
| 1202 |
+
{
|
| 1203 |
+
"epoch": 1.4273789649415694,
|
| 1204 |
+
"grad_norm": 1.2260764837265015,
|
| 1205 |
+
"learning_rate": 1.7459602925053494e-05,
|
| 1206 |
+
"loss": 0.5001,
|
| 1207 |
+
"step": 855
|
| 1208 |
+
},
|
| 1209 |
+
{
|
| 1210 |
+
"epoch": 1.4357262103505843,
|
| 1211 |
+
"grad_norm": 1.0423624515533447,
|
| 1212 |
+
"learning_rate": 1.7323337687932226e-05,
|
| 1213 |
+
"loss": 0.458,
|
| 1214 |
+
"step": 860
|
| 1215 |
+
},
|
| 1216 |
+
{
|
| 1217 |
+
"epoch": 1.4440734557595993,
|
| 1218 |
+
"grad_norm": 1.0051562786102295,
|
| 1219 |
+
"learning_rate": 1.7186875715650176e-05,
|
| 1220 |
+
"loss": 0.5584,
|
| 1221 |
+
"step": 865
|
| 1222 |
+
},
|
| 1223 |
+
{
|
| 1224 |
+
"epoch": 1.4524207011686143,
|
| 1225 |
+
"grad_norm": 1.1307841539382935,
|
| 1226 |
+
"learning_rate": 1.705022856350909e-05,
|
| 1227 |
+
"loss": 0.4934,
|
| 1228 |
+
"step": 870
|
| 1229 |
+
},
|
| 1230 |
+
{
|
| 1231 |
+
"epoch": 1.4607679465776293,
|
| 1232 |
+
"grad_norm": 1.3129503726959229,
|
| 1233 |
+
"learning_rate": 1.6913407802491346e-05,
|
| 1234 |
+
"loss": 0.5019,
|
| 1235 |
+
"step": 875
|
| 1236 |
+
},
|
| 1237 |
+
{
|
| 1238 |
+
"epoch": 1.4691151919866443,
|
| 1239 |
+
"grad_norm": 1.2901132106781006,
|
| 1240 |
+
"learning_rate": 1.677642501828015e-05,
|
| 1241 |
+
"loss": 0.4795,
|
| 1242 |
+
"step": 880
|
| 1243 |
+
},
|
| 1244 |
+
{
|
| 1245 |
+
"epoch": 1.4774624373956593,
|
| 1246 |
+
"grad_norm": 1.1699358224868774,
|
| 1247 |
+
"learning_rate": 1.663929181027846e-05,
|
| 1248 |
+
"loss": 0.4882,
|
| 1249 |
+
"step": 885
|
| 1250 |
+
},
|
| 1251 |
+
{
|
| 1252 |
+
"epoch": 1.4858096828046745,
|
| 1253 |
+
"grad_norm": 1.0698671340942383,
|
| 1254 |
+
"learning_rate": 1.6502019790626815e-05,
|
| 1255 |
+
"loss": 0.4892,
|
| 1256 |
+
"step": 890
|
| 1257 |
+
},
|
| 1258 |
+
{
|
| 1259 |
+
"epoch": 1.4941569282136895,
|
| 1260 |
+
"grad_norm": 1.1138356924057007,
|
| 1261 |
+
"learning_rate": 1.6364620583220006e-05,
|
| 1262 |
+
"loss": 0.4883,
|
| 1263 |
+
"step": 895
|
| 1264 |
+
},
|
| 1265 |
+
{
|
| 1266 |
+
"epoch": 1.5025041736227045,
|
| 1267 |
+
"grad_norm": 1.201592206954956,
|
| 1268 |
+
"learning_rate": 1.6227105822722803e-05,
|
| 1269 |
+
"loss": 0.4806,
|
| 1270 |
+
"step": 900
|
| 1271 |
+
},
|
| 1272 |
+
{
|
| 1273 |
+
"epoch": 1.5108514190317195,
|
| 1274 |
+
"grad_norm": 0.9779531359672546,
|
| 1275 |
+
"learning_rate": 1.608948715358478e-05,
|
| 1276 |
+
"loss": 0.4801,
|
| 1277 |
+
"step": 905
|
| 1278 |
+
},
|
| 1279 |
+
{
|
| 1280 |
+
"epoch": 1.5191986644407347,
|
| 1281 |
+
"grad_norm": 1.0987868309020996,
|
| 1282 |
+
"learning_rate": 1.595177622905424e-05,
|
| 1283 |
+
"loss": 0.4801,
|
| 1284 |
+
"step": 910
|
| 1285 |
+
},
|
| 1286 |
+
{
|
| 1287 |
+
"epoch": 1.5275459098497497,
|
| 1288 |
+
"grad_norm": 1.1521284580230713,
|
| 1289 |
+
"learning_rate": 1.581398471019149e-05,
|
| 1290 |
+
"loss": 0.5053,
|
| 1291 |
+
"step": 915
|
| 1292 |
+
},
|
| 1293 |
+
{
|
| 1294 |
+
"epoch": 1.5358931552587647,
|
| 1295 |
+
"grad_norm": 1.0605182647705078,
|
| 1296 |
+
"learning_rate": 1.5676124264881385e-05,
|
| 1297 |
+
"loss": 0.4479,
|
| 1298 |
+
"step": 920
|
| 1299 |
+
},
|
| 1300 |
+
{
|
| 1301 |
+
"epoch": 1.5442404006677797,
|
| 1302 |
+
"grad_norm": 1.2151260375976562,
|
| 1303 |
+
"learning_rate": 1.5538206566845314e-05,
|
| 1304 |
+
"loss": 0.4919,
|
| 1305 |
+
"step": 925
|
| 1306 |
+
},
|
| 1307 |
+
{
|
| 1308 |
+
"epoch": 1.5525876460767947,
|
| 1309 |
+
"grad_norm": 1.41464102268219,
|
| 1310 |
+
"learning_rate": 1.5400243294652713e-05,
|
| 1311 |
+
"loss": 0.4858,
|
| 1312 |
+
"step": 930
|
| 1313 |
+
},
|
| 1314 |
+
{
|
| 1315 |
+
"epoch": 1.5609348914858097,
|
| 1316 |
+
"grad_norm": 1.2979049682617188,
|
| 1317 |
+
"learning_rate": 1.5262246130732136e-05,
|
| 1318 |
+
"loss": 0.4869,
|
| 1319 |
+
"step": 935
|
| 1320 |
+
},
|
| 1321 |
+
{
|
| 1322 |
+
"epoch": 1.5692821368948247,
|
| 1323 |
+
"grad_norm": 1.1215578317642212,
|
| 1324 |
+
"learning_rate": 1.5124226760381998e-05,
|
| 1325 |
+
"loss": 0.4404,
|
| 1326 |
+
"step": 940
|
| 1327 |
+
},
|
| 1328 |
+
{
|
| 1329 |
+
"epoch": 1.5776293823038396,
|
| 1330 |
+
"grad_norm": 1.0793694257736206,
|
| 1331 |
+
"learning_rate": 1.498619687078113e-05,
|
| 1332 |
+
"loss": 0.4543,
|
| 1333 |
+
"step": 945
|
| 1334 |
+
},
|
| 1335 |
+
{
|
| 1336 |
+
"epoch": 1.5859766277128546,
|
| 1337 |
+
"grad_norm": 1.1885238885879517,
|
| 1338 |
+
"learning_rate": 1.4848168149999095e-05,
|
| 1339 |
+
"loss": 0.4252,
|
| 1340 |
+
"step": 950
|
| 1341 |
+
},
|
| 1342 |
+
{
|
| 1343 |
+
"epoch": 1.5943238731218696,
|
| 1344 |
+
"grad_norm": 1.1655299663543701,
|
| 1345 |
+
"learning_rate": 1.471015228600649e-05,
|
| 1346 |
+
"loss": 0.4326,
|
| 1347 |
+
"step": 955
|
| 1348 |
+
},
|
| 1349 |
+
{
|
| 1350 |
+
"epoch": 1.6026711185308848,
|
| 1351 |
+
"grad_norm": 1.3131167888641357,
|
| 1352 |
+
"learning_rate": 1.457216096568522e-05,
|
| 1353 |
+
"loss": 0.4895,
|
| 1354 |
+
"step": 960
|
| 1355 |
+
},
|
| 1356 |
+
{
|
| 1357 |
+
"epoch": 1.6110183639398998,
|
| 1358 |
+
"grad_norm": 1.0055269002914429,
|
| 1359 |
+
"learning_rate": 1.4434205873838892e-05,
|
| 1360 |
+
"loss": 0.4771,
|
| 1361 |
+
"step": 965
|
| 1362 |
+
},
|
| 1363 |
+
{
|
| 1364 |
+
"epoch": 1.6193656093489148,
|
| 1365 |
+
"grad_norm": 1.107136607170105,
|
| 1366 |
+
"learning_rate": 1.4296298692203365e-05,
|
| 1367 |
+
"loss": 0.4256,
|
| 1368 |
+
"step": 970
|
| 1369 |
+
},
|
| 1370 |
+
{
|
| 1371 |
+
"epoch": 1.62771285475793,
|
| 1372 |
+
"grad_norm": 1.0414416790008545,
|
| 1373 |
+
"learning_rate": 1.415845109845755e-05,
|
| 1374 |
+
"loss": 0.4429,
|
| 1375 |
+
"step": 975
|
| 1376 |
+
},
|
| 1377 |
+
{
|
| 1378 |
+
"epoch": 1.636060100166945,
|
| 1379 |
+
"grad_norm": 1.153465747833252,
|
| 1380 |
+
"learning_rate": 1.402067476523462e-05,
|
| 1381 |
+
"loss": 0.4668,
|
| 1382 |
+
"step": 980
|
| 1383 |
+
},
|
| 1384 |
+
{
|
| 1385 |
+
"epoch": 1.64440734557596,
|
| 1386 |
+
"grad_norm": 1.2037125825881958,
|
| 1387 |
+
"learning_rate": 1.3882981359133525e-05,
|
| 1388 |
+
"loss": 0.5108,
|
| 1389 |
+
"step": 985
|
| 1390 |
+
},
|
| 1391 |
+
{
|
| 1392 |
+
"epoch": 1.652754590984975,
|
| 1393 |
+
"grad_norm": 1.0885024070739746,
|
| 1394 |
+
"learning_rate": 1.3745382539731154e-05,
|
| 1395 |
+
"loss": 0.4727,
|
| 1396 |
+
"step": 990
|
| 1397 |
+
},
|
| 1398 |
+
{
|
| 1399 |
+
"epoch": 1.66110183639399,
|
| 1400 |
+
"grad_norm": 1.2434452772140503,
|
| 1401 |
+
"learning_rate": 1.360788995859499e-05,
|
| 1402 |
+
"loss": 0.4383,
|
| 1403 |
+
"step": 995
|
| 1404 |
+
},
|
| 1405 |
+
{
|
| 1406 |
+
"epoch": 1.669449081803005,
|
| 1407 |
+
"grad_norm": 1.1903680562973022,
|
| 1408 |
+
"learning_rate": 1.3470515258296482e-05,
|
| 1409 |
+
"loss": 0.4834,
|
| 1410 |
+
"step": 1000
|
| 1411 |
+
},
|
| 1412 |
+
{
|
| 1413 |
+
"epoch": 1.67779632721202,
|
| 1414 |
+
"grad_norm": 1.2532165050506592,
|
| 1415 |
+
"learning_rate": 1.3333270071425214e-05,
|
| 1416 |
+
"loss": 0.4521,
|
| 1417 |
+
"step": 1005
|
| 1418 |
+
},
|
| 1419 |
+
{
|
| 1420 |
+
"epoch": 1.686143572621035,
|
| 1421 |
+
"grad_norm": 1.093607783317566,
|
| 1422 |
+
"learning_rate": 1.3196166019603833e-05,
|
| 1423 |
+
"loss": 0.4475,
|
| 1424 |
+
"step": 1010
|
| 1425 |
+
},
|
| 1426 |
+
{
|
| 1427 |
+
"epoch": 1.69449081803005,
|
| 1428 |
+
"grad_norm": 1.2197047472000122,
|
| 1429 |
+
"learning_rate": 1.3059214712503985e-05,
|
| 1430 |
+
"loss": 0.4395,
|
| 1431 |
+
"step": 1015
|
| 1432 |
+
},
|
| 1433 |
+
{
|
| 1434 |
+
"epoch": 1.702838063439065,
|
| 1435 |
+
"grad_norm": 1.1287281513214111,
|
| 1436 |
+
"learning_rate": 1.2922427746863223e-05,
|
| 1437 |
+
"loss": 0.4472,
|
| 1438 |
+
"step": 1020
|
| 1439 |
+
},
|
| 1440 |
+
{
|
| 1441 |
+
"epoch": 1.7111853088480802,
|
| 1442 |
+
"grad_norm": 1.238547921180725,
|
| 1443 |
+
"learning_rate": 1.2785816705503031e-05,
|
| 1444 |
+
"loss": 0.454,
|
| 1445 |
+
"step": 1025
|
| 1446 |
+
},
|
| 1447 |
+
{
|
| 1448 |
+
"epoch": 1.7195325542570952,
|
| 1449 |
+
"grad_norm": 1.0695170164108276,
|
| 1450 |
+
"learning_rate": 1.2649393156348005e-05,
|
| 1451 |
+
"loss": 0.4189,
|
| 1452 |
+
"step": 1030
|
| 1453 |
+
},
|
| 1454 |
+
{
|
| 1455 |
+
"epoch": 1.7278797996661102,
|
| 1456 |
+
"grad_norm": 1.3531237840652466,
|
| 1457 |
+
"learning_rate": 1.2513168651446318e-05,
|
| 1458 |
+
"loss": 0.439,
|
| 1459 |
+
"step": 1035
|
| 1460 |
+
},
|
| 1461 |
+
{
|
| 1462 |
+
"epoch": 1.7362270450751254,
|
| 1463 |
+
"grad_norm": 1.1776330471038818,
|
| 1464 |
+
"learning_rate": 1.2377154725991497e-05,
|
| 1465 |
+
"loss": 0.3994,
|
| 1466 |
+
"step": 1040
|
| 1467 |
+
},
|
| 1468 |
+
{
|
| 1469 |
+
"epoch": 1.7445742904841404,
|
| 1470 |
+
"grad_norm": 1.0121259689331055,
|
| 1471 |
+
"learning_rate": 1.2241362897345671e-05,
|
| 1472 |
+
"loss": 0.44,
|
| 1473 |
+
"step": 1045
|
| 1474 |
+
},
|
| 1475 |
+
{
|
| 1476 |
+
"epoch": 1.7529215358931554,
|
| 1477 |
+
"grad_norm": 1.1822965145111084,
|
| 1478 |
+
"learning_rate": 1.2105804664064287e-05,
|
| 1479 |
+
"loss": 0.4405,
|
| 1480 |
+
"step": 1050
|
| 1481 |
+
},
|
| 1482 |
+
{
|
| 1483 |
+
"epoch": 1.7612687813021703,
|
| 1484 |
+
"grad_norm": 1.242809534072876,
|
| 1485 |
+
"learning_rate": 1.1970491504922466e-05,
|
| 1486 |
+
"loss": 0.4397,
|
| 1487 |
+
"step": 1055
|
| 1488 |
+
},
|
| 1489 |
+
{
|
| 1490 |
+
"epoch": 1.7696160267111853,
|
| 1491 |
+
"grad_norm": 1.1859283447265625,
|
| 1492 |
+
"learning_rate": 1.1835434877942963e-05,
|
| 1493 |
+
"loss": 0.4191,
|
| 1494 |
+
"step": 1060
|
| 1495 |
+
},
|
| 1496 |
+
{
|
| 1497 |
+
"epoch": 1.7779632721202003,
|
| 1498 |
+
"grad_norm": 1.1887779235839844,
|
| 1499 |
+
"learning_rate": 1.170064621942597e-05,
|
| 1500 |
+
"loss": 0.4529,
|
| 1501 |
+
"step": 1065
|
| 1502 |
+
},
|
| 1503 |
+
{
|
| 1504 |
+
"epoch": 1.7863105175292153,
|
| 1505 |
+
"grad_norm": 1.1291053295135498,
|
| 1506 |
+
"learning_rate": 1.1566136942980678e-05,
|
| 1507 |
+
"loss": 0.4073,
|
| 1508 |
+
"step": 1070
|
| 1509 |
+
},
|
| 1510 |
+
{
|
| 1511 |
+
"epoch": 1.7946577629382303,
|
| 1512 |
+
"grad_norm": 1.1099392175674438,
|
| 1513 |
+
"learning_rate": 1.1431918438558826e-05,
|
| 1514 |
+
"loss": 0.3946,
|
| 1515 |
+
"step": 1075
|
| 1516 |
+
},
|
| 1517 |
+
{
|
| 1518 |
+
"epoch": 1.8030050083472453,
|
| 1519 |
+
"grad_norm": 1.2921284437179565,
|
| 1520 |
+
"learning_rate": 1.1298002071490197e-05,
|
| 1521 |
+
"loss": 0.462,
|
| 1522 |
+
"step": 1080
|
| 1523 |
+
},
|
| 1524 |
+
{
|
| 1525 |
+
"epoch": 1.8113522537562603,
|
| 1526 |
+
"grad_norm": 1.225915551185608,
|
| 1527 |
+
"learning_rate": 1.1164399181520274e-05,
|
| 1528 |
+
"loss": 0.3957,
|
| 1529 |
+
"step": 1085
|
| 1530 |
+
},
|
| 1531 |
+
{
|
| 1532 |
+
"epoch": 1.8196994991652755,
|
| 1533 |
+
"grad_norm": 1.2412129640579224,
|
| 1534 |
+
"learning_rate": 1.1031121081849953e-05,
|
| 1535 |
+
"loss": 0.4259,
|
| 1536 |
+
"step": 1090
|
| 1537 |
+
},
|
| 1538 |
+
{
|
| 1539 |
+
"epoch": 1.8280467445742905,
|
| 1540 |
+
"grad_norm": 1.3541712760925293,
|
| 1541 |
+
"learning_rate": 1.0898179058177607e-05,
|
| 1542 |
+
"loss": 0.4039,
|
| 1543 |
+
"step": 1095
|
| 1544 |
+
},
|
| 1545 |
+
{
|
| 1546 |
+
"epoch": 1.8363939899833055,
|
| 1547 |
+
"grad_norm": 1.2248936891555786,
|
| 1548 |
+
"learning_rate": 1.0765584367743439e-05,
|
| 1549 |
+
"loss": 0.4267,
|
| 1550 |
+
"step": 1100
|
| 1551 |
+
},
|
| 1552 |
+
{
|
| 1553 |
+
"epoch": 1.8447412353923205,
|
| 1554 |
+
"grad_norm": 1.2510377168655396,
|
| 1555 |
+
"learning_rate": 1.0633348238376206e-05,
|
| 1556 |
+
"loss": 0.433,
|
| 1557 |
+
"step": 1105
|
| 1558 |
+
},
|
| 1559 |
+
{
|
| 1560 |
+
"epoch": 1.8530884808013357,
|
| 1561 |
+
"grad_norm": 1.373317837715149,
|
| 1562 |
+
"learning_rate": 1.0501481867542533e-05,
|
| 1563 |
+
"loss": 0.4132,
|
| 1564 |
+
"step": 1110
|
| 1565 |
+
},
|
| 1566 |
+
{
|
| 1567 |
+
"epoch": 1.8614357262103507,
|
| 1568 |
+
"grad_norm": 1.201568603515625,
|
| 1569 |
+
"learning_rate": 1.0369996421398667e-05,
|
| 1570 |
+
"loss": 0.4068,
|
| 1571 |
+
"step": 1115
|
| 1572 |
+
},
|
| 1573 |
+
{
|
| 1574 |
+
"epoch": 1.8697829716193657,
|
| 1575 |
+
"grad_norm": 1.2778894901275635,
|
| 1576 |
+
"learning_rate": 1.0238903033845003e-05,
|
| 1577 |
+
"loss": 0.413,
|
| 1578 |
+
"step": 1120
|
| 1579 |
+
},
|
| 1580 |
+
{
|
| 1581 |
+
"epoch": 1.8781302170283807,
|
| 1582 |
+
"grad_norm": 1.121774673461914,
|
| 1583 |
+
"learning_rate": 1.0108212805583262e-05,
|
| 1584 |
+
"loss": 0.453,
|
| 1585 |
+
"step": 1125
|
| 1586 |
+
},
|
| 1587 |
+
{
|
| 1588 |
+
"epoch": 1.8864774624373957,
|
| 1589 |
+
"grad_norm": 1.190560221672058,
|
| 1590 |
+
"learning_rate": 9.977936803176514e-06,
|
| 1591 |
+
"loss": 0.3943,
|
| 1592 |
+
"step": 1130
|
| 1593 |
+
},
|
| 1594 |
+
{
|
| 1595 |
+
"epoch": 1.8948247078464107,
|
| 1596 |
+
"grad_norm": 1.1311967372894287,
|
| 1597 |
+
"learning_rate": 9.848086058112108e-06,
|
| 1598 |
+
"loss": 0.3967,
|
| 1599 |
+
"step": 1135
|
| 1600 |
+
},
|
| 1601 |
+
{
|
| 1602 |
+
"epoch": 1.9031719532554257,
|
| 1603 |
+
"grad_norm": 1.1045544147491455,
|
| 1604 |
+
"learning_rate": 9.718671565867509e-06,
|
| 1605 |
+
"loss": 0.4481,
|
| 1606 |
+
"step": 1140
|
| 1607 |
+
},
|
| 1608 |
+
{
|
| 1609 |
+
"epoch": 1.9115191986644406,
|
| 1610 |
+
"grad_norm": 1.2753934860229492,
|
| 1611 |
+
"learning_rate": 9.589704284979256e-06,
|
| 1612 |
+
"loss": 0.3848,
|
| 1613 |
+
"step": 1145
|
| 1614 |
+
},
|
| 1615 |
+
{
|
| 1616 |
+
"epoch": 1.9198664440734556,
|
| 1617 |
+
"grad_norm": 1.1316887140274048,
|
| 1618 |
+
"learning_rate": 9.461195136115004e-06,
|
| 1619 |
+
"loss": 0.3999,
|
| 1620 |
+
"step": 1150
|
| 1621 |
+
},
|
| 1622 |
+
{
|
| 1623 |
+
"epoch": 1.9282136894824706,
|
| 1624 |
+
"grad_norm": 1.0906206369400024,
|
| 1625 |
+
"learning_rate": 9.333155001148784e-06,
|
| 1626 |
+
"loss": 0.3698,
|
| 1627 |
+
"step": 1155
|
| 1628 |
+
},
|
| 1629 |
+
{
|
| 1630 |
+
"epoch": 1.9365609348914858,
|
| 1631 |
+
"grad_norm": 1.237143635749817,
|
| 1632 |
+
"learning_rate": 9.20559472223957e-06,
|
| 1633 |
+
"loss": 0.3911,
|
| 1634 |
+
"step": 1160
|
| 1635 |
+
},
|
| 1636 |
+
{
|
| 1637 |
+
"epoch": 1.9449081803005008,
|
| 1638 |
+
"grad_norm": 1.1726173162460327,
|
| 1639 |
+
"learning_rate": 9.07852510091316e-06,
|
| 1640 |
+
"loss": 0.4138,
|
| 1641 |
+
"step": 1165
|
| 1642 |
+
},
|
| 1643 |
+
{
|
| 1644 |
+
"epoch": 1.9532554257095158,
|
| 1645 |
+
"grad_norm": 1.216637134552002,
|
| 1646 |
+
"learning_rate": 8.951956897147543e-06,
|
| 1647 |
+
"loss": 0.3804,
|
| 1648 |
+
"step": 1170
|
| 1649 |
+
},
|
| 1650 |
+
{
|
| 1651 |
+
"epoch": 1.961602671118531,
|
| 1652 |
+
"grad_norm": 1.3176758289337158,
|
| 1653 |
+
"learning_rate": 8.825900828461756e-06,
|
| 1654 |
+
"loss": 0.4488,
|
| 1655 |
+
"step": 1175
|
| 1656 |
+
},
|
| 1657 |
+
{
|
| 1658 |
+
"epoch": 1.969949916527546,
|
| 1659 |
+
"grad_norm": 1.130367398262024,
|
| 1660 |
+
"learning_rate": 8.700367569008377e-06,
|
| 1661 |
+
"loss": 0.406,
|
| 1662 |
+
"step": 1180
|
| 1663 |
+
},
|
| 1664 |
+
{
|
| 1665 |
+
"epoch": 1.978297161936561,
|
| 1666 |
+
"grad_norm": 1.182192087173462,
|
| 1667 |
+
"learning_rate": 8.575367748669616e-06,
|
| 1668 |
+
"loss": 0.3782,
|
| 1669 |
+
"step": 1185
|
| 1670 |
+
},
|
| 1671 |
+
{
|
| 1672 |
+
"epoch": 1.986644407345576,
|
| 1673 |
+
"grad_norm": 1.3355133533477783,
|
| 1674 |
+
"learning_rate": 8.450911952157242e-06,
|
| 1675 |
+
"loss": 0.441,
|
| 1676 |
+
"step": 1190
|
| 1677 |
+
},
|
| 1678 |
+
{
|
| 1679 |
+
"epoch": 1.994991652754591,
|
| 1680 |
+
"grad_norm": 1.32088041305542,
|
| 1681 |
+
"learning_rate": 8.327010718116263e-06,
|
| 1682 |
+
"loss": 0.381,
|
| 1683 |
+
"step": 1195
|
| 1684 |
+
},
|
| 1685 |
+
{
|
| 1686 |
+
"epoch": 2.003338898163606,
|
| 1687 |
+
"grad_norm": 1.1774063110351562,
|
| 1688 |
+
"learning_rate": 8.203674538232566e-06,
|
| 1689 |
+
"loss": 0.3642,
|
| 1690 |
+
"step": 1200
|
| 1691 |
+
},
|
| 1692 |
+
{
|
| 1693 |
+
"epoch": 2.011686143572621,
|
| 1694 |
+
"grad_norm": 1.1964908838272095,
|
| 1695 |
+
"learning_rate": 8.080913856344456e-06,
|
| 1696 |
+
"loss": 0.368,
|
| 1697 |
+
"step": 1205
|
| 1698 |
+
},
|
| 1699 |
+
{
|
| 1700 |
+
"epoch": 2.020033388981636,
|
| 1701 |
+
"grad_norm": 1.349615216255188,
|
| 1702 |
+
"learning_rate": 7.958739067558347e-06,
|
| 1703 |
+
"loss": 0.3084,
|
| 1704 |
+
"step": 1210
|
| 1705 |
+
},
|
| 1706 |
+
{
|
| 1707 |
+
"epoch": 2.028380634390651,
|
| 1708 |
+
"grad_norm": 1.2166770696640015,
|
| 1709 |
+
"learning_rate": 7.837160517368504e-06,
|
| 1710 |
+
"loss": 0.3411,
|
| 1711 |
+
"step": 1215
|
| 1712 |
+
},
|
| 1713 |
+
{
|
| 1714 |
+
"epoch": 2.036727879799666,
|
| 1715 |
+
"grad_norm": 1.169980525970459,
|
| 1716 |
+
"learning_rate": 7.716188500781008e-06,
|
| 1717 |
+
"loss": 0.3292,
|
| 1718 |
+
"step": 1220
|
| 1719 |
+
},
|
| 1720 |
+
{
|
| 1721 |
+
"epoch": 2.045075125208681,
|
| 1722 |
+
"grad_norm": 1.3812862634658813,
|
| 1723 |
+
"learning_rate": 7.595833261441987e-06,
|
| 1724 |
+
"loss": 0.3652,
|
| 1725 |
+
"step": 1225
|
| 1726 |
+
},
|
| 1727 |
+
{
|
| 1728 |
+
"epoch": 2.053422370617696,
|
| 1729 |
+
"grad_norm": 1.1697107553482056,
|
| 1730 |
+
"learning_rate": 7.476104990770234e-06,
|
| 1731 |
+
"loss": 0.3319,
|
| 1732 |
+
"step": 1230
|
| 1733 |
+
},
|
| 1734 |
+
{
|
| 1735 |
+
"epoch": 2.0617696160267114,
|
| 1736 |
+
"grad_norm": 1.1110737323760986,
|
| 1737 |
+
"learning_rate": 7.3570138270942006e-06,
|
| 1738 |
+
"loss": 0.3422,
|
| 1739 |
+
"step": 1235
|
| 1740 |
+
},
|
| 1741 |
+
{
|
| 1742 |
+
"epoch": 2.0701168614357264,
|
| 1743 |
+
"grad_norm": 1.0641626119613647,
|
| 1744 |
+
"learning_rate": 7.238569854793507e-06,
|
| 1745 |
+
"loss": 0.2898,
|
| 1746 |
+
"step": 1240
|
| 1747 |
+
},
|
| 1748 |
+
{
|
| 1749 |
+
"epoch": 2.0784641068447414,
|
| 1750 |
+
"grad_norm": 1.2949285507202148,
|
| 1751 |
+
"learning_rate": 7.1207831034450284e-06,
|
| 1752 |
+
"loss": 0.3083,
|
| 1753 |
+
"step": 1245
|
| 1754 |
+
},
|
| 1755 |
+
{
|
| 1756 |
+
"epoch": 2.0868113522537564,
|
| 1757 |
+
"grad_norm": 1.298567295074463,
|
| 1758 |
+
"learning_rate": 7.003663546973589e-06,
|
| 1759 |
+
"loss": 0.3297,
|
| 1760 |
+
"step": 1250
|
| 1761 |
+
},
|
| 1762 |
+
{
|
| 1763 |
+
"epoch": 2.0951585976627713,
|
| 1764 |
+
"grad_norm": 1.1909589767456055,
|
| 1765 |
+
"learning_rate": 6.887221102807429e-06,
|
| 1766 |
+
"loss": 0.3434,
|
| 1767 |
+
"step": 1255
|
| 1768 |
+
},
|
| 1769 |
+
{
|
| 1770 |
+
"epoch": 2.1035058430717863,
|
| 1771 |
+
"grad_norm": 1.1845543384552002,
|
| 1772 |
+
"learning_rate": 6.771465631038372e-06,
|
| 1773 |
+
"loss": 0.3456,
|
| 1774 |
+
"step": 1260
|
| 1775 |
+
},
|
| 1776 |
+
{
|
| 1777 |
+
"epoch": 2.1118530884808013,
|
| 1778 |
+
"grad_norm": 1.107622504234314,
|
| 1779 |
+
"learning_rate": 6.656406933586956e-06,
|
| 1780 |
+
"loss": 0.2788,
|
| 1781 |
+
"step": 1265
|
| 1782 |
+
},
|
| 1783 |
+
{
|
| 1784 |
+
"epoch": 2.1202003338898163,
|
| 1785 |
+
"grad_norm": 1.2579939365386963,
|
| 1786 |
+
"learning_rate": 6.542054753372353e-06,
|
| 1787 |
+
"loss": 0.3254,
|
| 1788 |
+
"step": 1270
|
| 1789 |
+
},
|
| 1790 |
+
{
|
| 1791 |
+
"epoch": 2.1285475792988313,
|
| 1792 |
+
"grad_norm": 1.205032467842102,
|
| 1793 |
+
"learning_rate": 6.428418773487424e-06,
|
| 1794 |
+
"loss": 0.3317,
|
| 1795 |
+
"step": 1275
|
| 1796 |
+
},
|
| 1797 |
+
{
|
| 1798 |
+
"epoch": 2.1368948247078463,
|
| 1799 |
+
"grad_norm": 1.1548906564712524,
|
| 1800 |
+
"learning_rate": 6.315508616378746e-06,
|
| 1801 |
+
"loss": 0.3284,
|
| 1802 |
+
"step": 1280
|
| 1803 |
+
},
|
| 1804 |
+
{
|
| 1805 |
+
"epoch": 2.1452420701168613,
|
| 1806 |
+
"grad_norm": 1.0587049722671509,
|
| 1807 |
+
"learning_rate": 6.2033338430317955e-06,
|
| 1808 |
+
"loss": 0.3311,
|
| 1809 |
+
"step": 1285
|
| 1810 |
+
},
|
| 1811 |
+
{
|
| 1812 |
+
"epoch": 2.1535893155258763,
|
| 1813 |
+
"grad_norm": 1.066192388534546,
|
| 1814 |
+
"learning_rate": 6.09190395216137e-06,
|
| 1815 |
+
"loss": 0.3281,
|
| 1816 |
+
"step": 1290
|
| 1817 |
+
},
|
| 1818 |
+
{
|
| 1819 |
+
"epoch": 2.1619365609348913,
|
| 1820 |
+
"grad_norm": 1.2256226539611816,
|
| 1821 |
+
"learning_rate": 5.981228379407247e-06,
|
| 1822 |
+
"loss": 0.3081,
|
| 1823 |
+
"step": 1295
|
| 1824 |
+
},
|
| 1825 |
+
{
|
| 1826 |
+
"epoch": 2.1702838063439067,
|
| 1827 |
+
"grad_norm": 1.1922575235366821,
|
| 1828 |
+
"learning_rate": 5.871316496535195e-06,
|
| 1829 |
+
"loss": 0.3078,
|
| 1830 |
+
"step": 1300
|
| 1831 |
+
},
|
| 1832 |
+
{
|
| 1833 |
+
"epoch": 2.1786310517529217,
|
| 1834 |
+
"grad_norm": 1.2795957326889038,
|
| 1835 |
+
"learning_rate": 5.762177610643374e-06,
|
| 1836 |
+
"loss": 0.3416,
|
| 1837 |
+
"step": 1305
|
| 1838 |
+
},
|
| 1839 |
+
{
|
| 1840 |
+
"epoch": 2.1869782971619367,
|
| 1841 |
+
"grad_norm": 1.1221518516540527,
|
| 1842 |
+
"learning_rate": 5.653820963374267e-06,
|
| 1843 |
+
"loss": 0.3337,
|
| 1844 |
+
"step": 1310
|
| 1845 |
+
},
|
| 1846 |
+
{
|
| 1847 |
+
"epoch": 2.1953255425709517,
|
| 1848 |
+
"grad_norm": 1.1923211812973022,
|
| 1849 |
+
"learning_rate": 5.546255730132095e-06,
|
| 1850 |
+
"loss": 0.3098,
|
| 1851 |
+
"step": 1315
|
| 1852 |
+
},
|
| 1853 |
+
{
|
| 1854 |
+
"epoch": 2.2036727879799667,
|
| 1855 |
+
"grad_norm": 1.3422815799713135,
|
| 1856 |
+
"learning_rate": 5.439491019305859e-06,
|
| 1857 |
+
"loss": 0.2797,
|
| 1858 |
+
"step": 1320
|
| 1859 |
+
},
|
| 1860 |
+
{
|
| 1861 |
+
"epoch": 2.2120200333889817,
|
| 1862 |
+
"grad_norm": 1.1795589923858643,
|
| 1863 |
+
"learning_rate": 5.333535871498089e-06,
|
| 1864 |
+
"loss": 0.3379,
|
| 1865 |
+
"step": 1325
|
| 1866 |
+
},
|
| 1867 |
+
{
|
| 1868 |
+
"epoch": 2.2203672787979967,
|
| 1869 |
+
"grad_norm": 1.2858433723449707,
|
| 1870 |
+
"learning_rate": 5.2283992587592594e-06,
|
| 1871 |
+
"loss": 0.3108,
|
| 1872 |
+
"step": 1330
|
| 1873 |
+
},
|
| 1874 |
+
{
|
| 1875 |
+
"epoch": 2.2287145242070117,
|
| 1876 |
+
"grad_norm": 1.1089833974838257,
|
| 1877 |
+
"learning_rate": 5.1240900838281106e-06,
|
| 1878 |
+
"loss": 0.3379,
|
| 1879 |
+
"step": 1335
|
| 1880 |
+
},
|
| 1881 |
+
{
|
| 1882 |
+
"epoch": 2.2370617696160267,
|
| 1883 |
+
"grad_norm": 1.135477066040039,
|
| 1884 |
+
"learning_rate": 5.02061717937773e-06,
|
| 1885 |
+
"loss": 0.3071,
|
| 1886 |
+
"step": 1340
|
| 1887 |
+
},
|
| 1888 |
+
{
|
| 1889 |
+
"epoch": 2.2454090150250416,
|
| 1890 |
+
"grad_norm": 1.2192907333374023,
|
| 1891 |
+
"learning_rate": 4.917989307267676e-06,
|
| 1892 |
+
"loss": 0.3286,
|
| 1893 |
+
"step": 1345
|
| 1894 |
+
},
|
| 1895 |
+
{
|
| 1896 |
+
"epoch": 2.2537562604340566,
|
| 1897 |
+
"grad_norm": 1.0953881740570068,
|
| 1898 |
+
"learning_rate": 4.816215157801991e-06,
|
| 1899 |
+
"loss": 0.3088,
|
| 1900 |
+
"step": 1350
|
| 1901 |
+
},
|
| 1902 |
+
{
|
| 1903 |
+
"epoch": 2.2621035058430716,
|
| 1904 |
+
"grad_norm": 1.331865668296814,
|
| 1905 |
+
"learning_rate": 4.715303348993358e-06,
|
| 1906 |
+
"loss": 0.3181,
|
| 1907 |
+
"step": 1355
|
| 1908 |
+
},
|
| 1909 |
+
{
|
| 1910 |
+
"epoch": 2.2704507512520866,
|
| 1911 |
+
"grad_norm": 1.184768557548523,
|
| 1912 |
+
"learning_rate": 4.615262425833342e-06,
|
| 1913 |
+
"loss": 0.3033,
|
| 1914 |
+
"step": 1360
|
| 1915 |
+
},
|
| 1916 |
+
{
|
| 1917 |
+
"epoch": 2.278797996661102,
|
| 1918 |
+
"grad_norm": 1.2596111297607422,
|
| 1919 |
+
"learning_rate": 4.516100859568791e-06,
|
| 1920 |
+
"loss": 0.3133,
|
| 1921 |
+
"step": 1365
|
| 1922 |
+
},
|
| 1923 |
+
{
|
| 1924 |
+
"epoch": 2.287145242070117,
|
| 1925 |
+
"grad_norm": 1.2273879051208496,
|
| 1926 |
+
"learning_rate": 4.41782704698457e-06,
|
| 1927 |
+
"loss": 0.2792,
|
| 1928 |
+
"step": 1370
|
| 1929 |
+
},
|
| 1930 |
+
{
|
| 1931 |
+
"epoch": 2.295492487479132,
|
| 1932 |
+
"grad_norm": 1.301862120628357,
|
| 1933 |
+
"learning_rate": 4.320449309692462e-06,
|
| 1934 |
+
"loss": 0.3071,
|
| 1935 |
+
"step": 1375
|
| 1936 |
+
},
|
| 1937 |
+
{
|
| 1938 |
+
"epoch": 2.303839732888147,
|
| 1939 |
+
"grad_norm": 1.1422520875930786,
|
| 1940 |
+
"learning_rate": 4.223975893426577e-06,
|
| 1941 |
+
"loss": 0.3188,
|
| 1942 |
+
"step": 1380
|
| 1943 |
+
},
|
| 1944 |
+
{
|
| 1945 |
+
"epoch": 2.312186978297162,
|
| 1946 |
+
"grad_norm": 1.424318790435791,
|
| 1947 |
+
"learning_rate": 4.12841496734509e-06,
|
| 1948 |
+
"loss": 0.3307,
|
| 1949 |
+
"step": 1385
|
| 1950 |
+
},
|
| 1951 |
+
{
|
| 1952 |
+
"epoch": 2.320534223706177,
|
| 1953 |
+
"grad_norm": 1.1943734884262085,
|
| 1954 |
+
"learning_rate": 4.033774623338489e-06,
|
| 1955 |
+
"loss": 0.3062,
|
| 1956 |
+
"step": 1390
|
| 1957 |
+
},
|
| 1958 |
+
{
|
| 1959 |
+
"epoch": 2.328881469115192,
|
| 1960 |
+
"grad_norm": 1.1429376602172852,
|
| 1961 |
+
"learning_rate": 3.940062875344395e-06,
|
| 1962 |
+
"loss": 0.3121,
|
| 1963 |
+
"step": 1395
|
| 1964 |
+
},
|
| 1965 |
+
{
|
| 1966 |
+
"epoch": 2.337228714524207,
|
| 1967 |
+
"grad_norm": 1.1322011947631836,
|
| 1968 |
+
"learning_rate": 3.847287658668945e-06,
|
| 1969 |
+
"loss": 0.3653,
|
| 1970 |
+
"step": 1400
|
| 1971 |
+
},
|
| 1972 |
+
{
|
| 1973 |
+
"epoch": 2.345575959933222,
|
| 1974 |
+
"grad_norm": 1.169260025024414,
|
| 1975 |
+
"learning_rate": 3.7554568293148496e-06,
|
| 1976 |
+
"loss": 0.315,
|
| 1977 |
+
"step": 1405
|
| 1978 |
+
},
|
| 1979 |
+
{
|
| 1980 |
+
"epoch": 2.353923205342237,
|
| 1981 |
+
"grad_norm": 1.1403430700302124,
|
| 1982 |
+
"learning_rate": 3.6645781633161556e-06,
|
| 1983 |
+
"loss": 0.2979,
|
| 1984 |
+
"step": 1410
|
| 1985 |
+
},
|
| 1986 |
+
{
|
| 1987 |
+
"epoch": 2.362270450751252,
|
| 1988 |
+
"grad_norm": 1.2582240104675293,
|
| 1989 |
+
"learning_rate": 3.5746593560798064e-06,
|
| 1990 |
+
"loss": 0.2999,
|
| 1991 |
+
"step": 1415
|
| 1992 |
+
},
|
| 1993 |
+
{
|
| 1994 |
+
"epoch": 2.370617696160267,
|
| 1995 |
+
"grad_norm": 1.211690068244934,
|
| 1996 |
+
"learning_rate": 3.485708021733999e-06,
|
| 1997 |
+
"loss": 0.3068,
|
| 1998 |
+
"step": 1420
|
| 1999 |
+
},
|
| 2000 |
+
{
|
| 2001 |
+
"epoch": 2.378964941569282,
|
| 2002 |
+
"grad_norm": 1.1625430583953857,
|
| 2003 |
+
"learning_rate": 3.3977316924834397e-06,
|
| 2004 |
+
"loss": 0.2964,
|
| 2005 |
+
"step": 1425
|
| 2006 |
+
},
|
| 2007 |
+
{
|
| 2008 |
+
"epoch": 2.3873121869782974,
|
| 2009 |
+
"grad_norm": 1.0662834644317627,
|
| 2010 |
+
"learning_rate": 3.3107378179715326e-06,
|
| 2011 |
+
"loss": 0.3357,
|
| 2012 |
+
"step": 1430
|
| 2013 |
+
},
|
| 2014 |
+
{
|
| 2015 |
+
"epoch": 2.395659432387312,
|
| 2016 |
+
"grad_norm": 1.2917218208312988,
|
| 2017 |
+
"learning_rate": 3.224733764649554e-06,
|
| 2018 |
+
"loss": 0.3361,
|
| 2019 |
+
"step": 1435
|
| 2020 |
+
},
|
| 2021 |
+
{
|
| 2022 |
+
"epoch": 2.4040066777963274,
|
| 2023 |
+
"grad_norm": 1.1366279125213623,
|
| 2024 |
+
"learning_rate": 3.1397268151528985e-06,
|
| 2025 |
+
"loss": 0.3132,
|
| 2026 |
+
"step": 1440
|
| 2027 |
+
},
|
| 2028 |
+
{
|
| 2029 |
+
"epoch": 2.4123539232053424,
|
| 2030 |
+
"grad_norm": 1.2154560089111328,
|
| 2031 |
+
"learning_rate": 3.055724167684362e-06,
|
| 2032 |
+
"loss": 0.3209,
|
| 2033 |
+
"step": 1445
|
| 2034 |
+
},
|
| 2035 |
+
{
|
| 2036 |
+
"epoch": 2.4207011686143574,
|
| 2037 |
+
"grad_norm": 1.1733973026275635,
|
| 2038 |
+
"learning_rate": 2.972732935404672e-06,
|
| 2039 |
+
"loss": 0.3223,
|
| 2040 |
+
"step": 1450
|
| 2041 |
+
},
|
| 2042 |
+
{
|
| 2043 |
+
"epoch": 2.4290484140233723,
|
| 2044 |
+
"grad_norm": 1.2234079837799072,
|
| 2045 |
+
"learning_rate": 2.890760145830099e-06,
|
| 2046 |
+
"loss": 0.3197,
|
| 2047 |
+
"step": 1455
|
| 2048 |
+
},
|
| 2049 |
+
{
|
| 2050 |
+
"epoch": 2.4373956594323873,
|
| 2051 |
+
"grad_norm": 1.095177173614502,
|
| 2052 |
+
"learning_rate": 2.809812740237419e-06,
|
| 2053 |
+
"loss": 0.2828,
|
| 2054 |
+
"step": 1460
|
| 2055 |
+
},
|
| 2056 |
+
{
|
| 2057 |
+
"epoch": 2.4457429048414023,
|
| 2058 |
+
"grad_norm": 1.2379978895187378,
|
| 2059 |
+
"learning_rate": 2.72989757307614e-06,
|
| 2060 |
+
"loss": 0.3119,
|
| 2061 |
+
"step": 1465
|
| 2062 |
+
},
|
| 2063 |
+
{
|
| 2064 |
+
"epoch": 2.4540901502504173,
|
| 2065 |
+
"grad_norm": 1.0995434522628784,
|
| 2066 |
+
"learning_rate": 2.6510214113880537e-06,
|
| 2067 |
+
"loss": 0.3007,
|
| 2068 |
+
"step": 1470
|
| 2069 |
+
},
|
| 2070 |
+
{
|
| 2071 |
+
"epoch": 2.4624373956594323,
|
| 2072 |
+
"grad_norm": 1.1604273319244385,
|
| 2073 |
+
"learning_rate": 2.573190934234268e-06,
|
| 2074 |
+
"loss": 0.289,
|
| 2075 |
+
"step": 1475
|
| 2076 |
+
},
|
| 2077 |
+
{
|
| 2078 |
+
"epoch": 2.4707846410684473,
|
| 2079 |
+
"grad_norm": 1.2481505870819092,
|
| 2080 |
+
"learning_rate": 2.4964127321295765e-06,
|
| 2081 |
+
"loss": 0.3036,
|
| 2082 |
+
"step": 1480
|
| 2083 |
+
},
|
| 2084 |
+
{
|
| 2085 |
+
"epoch": 2.4791318864774623,
|
| 2086 |
+
"grad_norm": 1.2154197692871094,
|
| 2087 |
+
"learning_rate": 2.4206933064844444e-06,
|
| 2088 |
+
"loss": 0.312,
|
| 2089 |
+
"step": 1485
|
| 2090 |
+
},
|
| 2091 |
+
{
|
| 2092 |
+
"epoch": 2.4874791318864773,
|
| 2093 |
+
"grad_norm": 1.003563642501831,
|
| 2094 |
+
"learning_rate": 2.3460390690544416e-06,
|
| 2095 |
+
"loss": 0.2781,
|
| 2096 |
+
"step": 1490
|
| 2097 |
+
},
|
| 2098 |
+
{
|
| 2099 |
+
"epoch": 2.4958263772954927,
|
| 2100 |
+
"grad_norm": 1.3014856576919556,
|
| 2101 |
+
"learning_rate": 2.2724563413973383e-06,
|
| 2102 |
+
"loss": 0.2908,
|
| 2103 |
+
"step": 1495
|
| 2104 |
+
},
|
| 2105 |
+
{
|
| 2106 |
+
"epoch": 2.5041736227045073,
|
| 2107 |
+
"grad_norm": 1.1999882459640503,
|
| 2108 |
+
"learning_rate": 2.1999513543377925e-06,
|
| 2109 |
+
"loss": 0.2833,
|
| 2110 |
+
"step": 1500
|
| 2111 |
+
},
|
| 2112 |
+
{
|
| 2113 |
+
"epoch": 2.5125208681135227,
|
| 2114 |
+
"grad_norm": 1.276909589767456,
|
| 2115 |
+
"learning_rate": 2.1285302474397433e-06,
|
| 2116 |
+
"loss": 0.2832,
|
| 2117 |
+
"step": 1505
|
| 2118 |
+
},
|
| 2119 |
+
{
|
| 2120 |
+
"epoch": 2.5208681135225377,
|
| 2121 |
+
"grad_norm": 1.2243189811706543,
|
| 2122 |
+
"learning_rate": 2.05819906848653e-06,
|
| 2123 |
+
"loss": 0.3162,
|
| 2124 |
+
"step": 1510
|
| 2125 |
+
},
|
| 2126 |
+
{
|
| 2127 |
+
"epoch": 2.5292153589315527,
|
| 2128 |
+
"grad_norm": 1.432253360748291,
|
| 2129 |
+
"learning_rate": 1.9889637729687594e-06,
|
| 2130 |
+
"loss": 0.3041,
|
| 2131 |
+
"step": 1515
|
| 2132 |
+
},
|
| 2133 |
+
{
|
| 2134 |
+
"epoch": 2.5375626043405677,
|
| 2135 |
+
"grad_norm": 1.238013744354248,
|
| 2136 |
+
"learning_rate": 1.920830223580042e-06,
|
| 2137 |
+
"loss": 0.2785,
|
| 2138 |
+
"step": 1520
|
| 2139 |
+
},
|
| 2140 |
+
{
|
| 2141 |
+
"epoch": 2.5459098497495827,
|
| 2142 |
+
"grad_norm": 1.1080212593078613,
|
| 2143 |
+
"learning_rate": 1.8538041897205133e-06,
|
| 2144 |
+
"loss": 0.3089,
|
| 2145 |
+
"step": 1525
|
| 2146 |
+
},
|
| 2147 |
+
{
|
| 2148 |
+
"epoch": 2.5542570951585977,
|
| 2149 |
+
"grad_norm": 1.309547781944275,
|
| 2150 |
+
"learning_rate": 1.787891347008334e-06,
|
| 2151 |
+
"loss": 0.3117,
|
| 2152 |
+
"step": 1530
|
| 2153 |
+
},
|
| 2154 |
+
{
|
| 2155 |
+
"epoch": 2.5626043405676127,
|
| 2156 |
+
"grad_norm": 1.40817391872406,
|
| 2157 |
+
"learning_rate": 1.7230972767990476e-06,
|
| 2158 |
+
"loss": 0.2868,
|
| 2159 |
+
"step": 1535
|
| 2160 |
+
},
|
| 2161 |
+
{
|
| 2162 |
+
"epoch": 2.5709515859766277,
|
| 2163 |
+
"grad_norm": 1.2554359436035156,
|
| 2164 |
+
"learning_rate": 1.6594274657130037e-06,
|
| 2165 |
+
"loss": 0.3251,
|
| 2166 |
+
"step": 1540
|
| 2167 |
+
},
|
| 2168 |
+
{
|
| 2169 |
+
"epoch": 2.5792988313856426,
|
| 2170 |
+
"grad_norm": 1.117342233657837,
|
| 2171 |
+
"learning_rate": 1.5968873051707372e-06,
|
| 2172 |
+
"loss": 0.2954,
|
| 2173 |
+
"step": 1545
|
| 2174 |
+
},
|
| 2175 |
+
{
|
| 2176 |
+
"epoch": 2.5876460767946576,
|
| 2177 |
+
"grad_norm": 1.2844130992889404,
|
| 2178 |
+
"learning_rate": 1.5354820909364354e-06,
|
| 2179 |
+
"loss": 0.3025,
|
| 2180 |
+
"step": 1550
|
| 2181 |
+
},
|
| 2182 |
+
{
|
| 2183 |
+
"epoch": 2.5959933222036726,
|
| 2184 |
+
"grad_norm": 1.1225556135177612,
|
| 2185 |
+
"learning_rate": 1.4752170226695294e-06,
|
| 2186 |
+
"loss": 0.3289,
|
| 2187 |
+
"step": 1555
|
| 2188 |
+
},
|
| 2189 |
+
{
|
| 2190 |
+
"epoch": 2.604340567612688,
|
| 2191 |
+
"grad_norm": 1.1798509359359741,
|
| 2192 |
+
"learning_rate": 1.4160972034843605e-06,
|
| 2193 |
+
"loss": 0.2835,
|
| 2194 |
+
"step": 1560
|
| 2195 |
+
},
|
| 2196 |
+
{
|
| 2197 |
+
"epoch": 2.6126878130217026,
|
| 2198 |
+
"grad_norm": 1.3689546585083008,
|
| 2199 |
+
"learning_rate": 1.358127639518097e-06,
|
| 2200 |
+
"loss": 0.3153,
|
| 2201 |
+
"step": 1565
|
| 2202 |
+
},
|
| 2203 |
+
{
|
| 2204 |
+
"epoch": 2.621035058430718,
|
| 2205 |
+
"grad_norm": 1.1938159465789795,
|
| 2206 |
+
"learning_rate": 1.3013132395067879e-06,
|
| 2207 |
+
"loss": 0.3093,
|
| 2208 |
+
"step": 1570
|
| 2209 |
+
},
|
| 2210 |
+
{
|
| 2211 |
+
"epoch": 2.629382303839733,
|
| 2212 |
+
"grad_norm": 1.2768663167953491,
|
| 2213 |
+
"learning_rate": 1.2456588143697378e-06,
|
| 2214 |
+
"loss": 0.2836,
|
| 2215 |
+
"step": 1575
|
| 2216 |
+
},
|
| 2217 |
+
{
|
| 2218 |
+
"epoch": 2.637729549248748,
|
| 2219 |
+
"grad_norm": 1.2760200500488281,
|
| 2220 |
+
"learning_rate": 1.1911690768021078e-06,
|
| 2221 |
+
"loss": 0.2843,
|
| 2222 |
+
"step": 1580
|
| 2223 |
+
},
|
| 2224 |
+
{
|
| 2225 |
+
"epoch": 2.646076794657763,
|
| 2226 |
+
"grad_norm": 1.3292090892791748,
|
| 2227 |
+
"learning_rate": 1.1378486408758587e-06,
|
| 2228 |
+
"loss": 0.303,
|
| 2229 |
+
"step": 1585
|
| 2230 |
+
},
|
| 2231 |
+
{
|
| 2232 |
+
"epoch": 2.654424040066778,
|
| 2233 |
+
"grad_norm": 1.1531957387924194,
|
| 2234 |
+
"learning_rate": 1.0857020216490433e-06,
|
| 2235 |
+
"loss": 0.2808,
|
| 2236 |
+
"step": 1590
|
| 2237 |
+
},
|
| 2238 |
+
{
|
| 2239 |
+
"epoch": 2.662771285475793,
|
| 2240 |
+
"grad_norm": 1.0699784755706787,
|
| 2241 |
+
"learning_rate": 1.0347336347834757e-06,
|
| 2242 |
+
"loss": 0.3131,
|
| 2243 |
+
"step": 1595
|
| 2244 |
+
},
|
| 2245 |
+
{
|
| 2246 |
+
"epoch": 2.671118530884808,
|
| 2247 |
+
"grad_norm": 1.2488611936569214,
|
| 2248 |
+
"learning_rate": 9.849477961708354e-07,
|
| 2249 |
+
"loss": 0.3096,
|
| 2250 |
+
"step": 1600
|
| 2251 |
+
},
|
| 2252 |
+
{
|
| 2253 |
+
"epoch": 2.679465776293823,
|
| 2254 |
+
"grad_norm": 1.283469796180725,
|
| 2255 |
+
"learning_rate": 9.363487215671862e-07,
|
| 2256 |
+
"loss": 0.3078,
|
| 2257 |
+
"step": 1605
|
| 2258 |
+
},
|
| 2259 |
+
{
|
| 2260 |
+
"epoch": 2.687813021702838,
|
| 2261 |
+
"grad_norm": 1.3012709617614746,
|
| 2262 |
+
"learning_rate": 8.889405262360267e-07,
|
| 2263 |
+
"loss": 0.3255,
|
| 2264 |
+
"step": 1610
|
| 2265 |
+
},
|
| 2266 |
+
{
|
| 2267 |
+
"epoch": 2.696160267111853,
|
| 2268 |
+
"grad_norm": 1.1437944173812866,
|
| 2269 |
+
"learning_rate": 8.427272245997797e-07,
|
| 2270 |
+
"loss": 0.3339,
|
| 2271 |
+
"step": 1615
|
| 2272 |
+
},
|
| 2273 |
+
{
|
| 2274 |
+
"epoch": 2.704507512520868,
|
| 2275 |
+
"grad_norm": 1.2112791538238525,
|
| 2276 |
+
"learning_rate": 7.97712729899891e-07,
|
| 2277 |
+
"loss": 0.3134,
|
| 2278 |
+
"step": 1620
|
| 2279 |
+
},
|
| 2280 |
+
{
|
| 2281 |
+
"epoch": 2.7128547579298834,
|
| 2282 |
+
"grad_norm": 1.1730989217758179,
|
| 2283 |
+
"learning_rate": 7.539008538654562e-07,
|
| 2284 |
+
"loss": 0.2974,
|
| 2285 |
+
"step": 1625
|
| 2286 |
+
},
|
| 2287 |
+
{
|
| 2288 |
+
"epoch": 2.721202003338898,
|
| 2289 |
+
"grad_norm": 1.1251872777938843,
|
| 2290 |
+
"learning_rate": 7.112953063904359e-07,
|
| 2291 |
+
"loss": 0.2864,
|
| 2292 |
+
"step": 1630
|
| 2293 |
+
},
|
| 2294 |
+
{
|
| 2295 |
+
"epoch": 2.7295492487479134,
|
| 2296 |
+
"grad_norm": 1.3749324083328247,
|
| 2297 |
+
"learning_rate": 6.698996952195396e-07,
|
| 2298 |
+
"loss": 0.2953,
|
| 2299 |
+
"step": 1635
|
| 2300 |
+
},
|
| 2301 |
+
{
|
| 2302 |
+
"epoch": 2.7378964941569284,
|
| 2303 |
+
"grad_norm": 1.1217886209487915,
|
| 2304 |
+
"learning_rate": 6.29717525642699e-07,
|
| 2305 |
+
"loss": 0.2705,
|
| 2306 |
+
"step": 1640
|
| 2307 |
+
},
|
| 2308 |
+
{
|
| 2309 |
+
"epoch": 2.7462437395659434,
|
| 2310 |
+
"grad_norm": 1.1644232273101807,
|
| 2311 |
+
"learning_rate": 5.907522001982718e-07,
|
| 2312 |
+
"loss": 0.2942,
|
| 2313 |
+
"step": 1645
|
| 2314 |
+
},
|
| 2315 |
+
{
|
| 2316 |
+
"epoch": 2.7545909849749584,
|
| 2317 |
+
"grad_norm": 1.1689884662628174,
|
| 2318 |
+
"learning_rate": 5.530070183849073e-07,
|
| 2319 |
+
"loss": 0.2773,
|
| 2320 |
+
"step": 1650
|
| 2321 |
+
},
|
| 2322 |
+
{
|
| 2323 |
+
"epoch": 2.7629382303839733,
|
| 2324 |
+
"grad_norm": 1.4353326559066772,
|
| 2325 |
+
"learning_rate": 5.164851763821566e-07,
|
| 2326 |
+
"loss": 0.2961,
|
| 2327 |
+
"step": 1655
|
| 2328 |
+
},
|
| 2329 |
+
{
|
| 2330 |
+
"epoch": 2.7712854757929883,
|
| 2331 |
+
"grad_norm": 1.1775834560394287,
|
| 2332 |
+
"learning_rate": 4.811897667798354e-07,
|
| 2333 |
+
"loss": 0.3062,
|
| 2334 |
+
"step": 1660
|
| 2335 |
+
},
|
| 2336 |
+
{
|
| 2337 |
+
"epoch": 2.7796327212020033,
|
| 2338 |
+
"grad_norm": 1.207299828529358,
|
| 2339 |
+
"learning_rate": 4.4712377831613084e-07,
|
| 2340 |
+
"loss": 0.3236,
|
| 2341 |
+
"step": 1665
|
| 2342 |
+
},
|
| 2343 |
+
{
|
| 2344 |
+
"epoch": 2.7879799666110183,
|
| 2345 |
+
"grad_norm": 1.1795037984848022,
|
| 2346 |
+
"learning_rate": 4.1429009562454324e-07,
|
| 2347 |
+
"loss": 0.2913,
|
| 2348 |
+
"step": 1670
|
| 2349 |
+
},
|
| 2350 |
+
{
|
| 2351 |
+
"epoch": 2.7963272120200333,
|
| 2352 |
+
"grad_norm": 1.2193044424057007,
|
| 2353 |
+
"learning_rate": 3.8269149898960355e-07,
|
| 2354 |
+
"loss": 0.3359,
|
| 2355 |
+
"step": 1675
|
| 2356 |
+
},
|
| 2357 |
+
{
|
| 2358 |
+
"epoch": 2.8046744574290483,
|
| 2359 |
+
"grad_norm": 1.1088463068008423,
|
| 2360 |
+
"learning_rate": 3.5233066411145985e-07,
|
| 2361 |
+
"loss": 0.2638,
|
| 2362 |
+
"step": 1680
|
| 2363 |
+
},
|
| 2364 |
+
{
|
| 2365 |
+
"epoch": 2.8130217028380633,
|
| 2366 |
+
"grad_norm": 1.12891685962677,
|
| 2367 |
+
"learning_rate": 3.232101618792954e-07,
|
| 2368 |
+
"loss": 0.3152,
|
| 2369 |
+
"step": 1685
|
| 2370 |
+
},
|
| 2371 |
+
{
|
| 2372 |
+
"epoch": 2.8213689482470787,
|
| 2373 |
+
"grad_norm": 1.2735636234283447,
|
| 2374 |
+
"learning_rate": 2.953324581536343e-07,
|
| 2375 |
+
"loss": 0.2737,
|
| 2376 |
+
"step": 1690
|
| 2377 |
+
},
|
| 2378 |
+
{
|
| 2379 |
+
"epoch": 2.8297161936560933,
|
| 2380 |
+
"grad_norm": 1.1320722103118896,
|
| 2381 |
+
"learning_rate": 2.686999135575402e-07,
|
| 2382 |
+
"loss": 0.3121,
|
| 2383 |
+
"step": 1695
|
| 2384 |
+
},
|
| 2385 |
+
{
|
| 2386 |
+
"epoch": 2.8380634390651087,
|
| 2387 |
+
"grad_norm": 1.2865349054336548,
|
| 2388 |
+
"learning_rate": 2.4331478327671297e-07,
|
| 2389 |
+
"loss": 0.3107,
|
| 2390 |
+
"step": 1700
|
| 2391 |
+
},
|
| 2392 |
+
{
|
| 2393 |
+
"epoch": 2.8464106844741233,
|
| 2394 |
+
"grad_norm": 1.2942535877227783,
|
| 2395 |
+
"learning_rate": 2.191792168685447e-07,
|
| 2396 |
+
"loss": 0.3138,
|
| 2397 |
+
"step": 1705
|
| 2398 |
+
},
|
| 2399 |
+
{
|
| 2400 |
+
"epoch": 2.8547579298831387,
|
| 2401 |
+
"grad_norm": 1.1381316184997559,
|
| 2402 |
+
"learning_rate": 1.962952580800753e-07,
|
| 2403 |
+
"loss": 0.3027,
|
| 2404 |
+
"step": 1710
|
| 2405 |
+
},
|
| 2406 |
+
{
|
| 2407 |
+
"epoch": 2.8631051752921537,
|
| 2408 |
+
"grad_norm": 1.1447514295578003,
|
| 2409 |
+
"learning_rate": 1.7466484467495937e-07,
|
| 2410 |
+
"loss": 0.2858,
|
| 2411 |
+
"step": 1715
|
| 2412 |
+
},
|
| 2413 |
+
{
|
| 2414 |
+
"epoch": 2.8714524207011687,
|
| 2415 |
+
"grad_norm": 1.1821335554122925,
|
| 2416 |
+
"learning_rate": 1.5428980826935558e-07,
|
| 2417 |
+
"loss": 0.2838,
|
| 2418 |
+
"step": 1720
|
| 2419 |
+
},
|
| 2420 |
+
{
|
| 2421 |
+
"epoch": 2.8797996661101837,
|
| 2422 |
+
"grad_norm": 1.1340967416763306,
|
| 2423 |
+
"learning_rate": 1.3517187417684262e-07,
|
| 2424 |
+
"loss": 0.2889,
|
| 2425 |
+
"step": 1725
|
| 2426 |
+
},
|
| 2427 |
+
{
|
| 2428 |
+
"epoch": 2.8881469115191987,
|
| 2429 |
+
"grad_norm": 1.3011823892593384,
|
| 2430 |
+
"learning_rate": 1.1731266126232265e-07,
|
| 2431 |
+
"loss": 0.2796,
|
| 2432 |
+
"step": 1730
|
| 2433 |
+
},
|
| 2434 |
+
{
|
| 2435 |
+
"epoch": 2.8964941569282137,
|
| 2436 |
+
"grad_norm": 1.1553633213043213,
|
| 2437 |
+
"learning_rate": 1.0071368180493756e-07,
|
| 2438 |
+
"loss": 0.3279,
|
| 2439 |
+
"step": 1735
|
| 2440 |
+
},
|
| 2441 |
+
{
|
| 2442 |
+
"epoch": 2.9048414023372287,
|
| 2443 |
+
"grad_norm": 1.1699466705322266,
|
| 2444 |
+
"learning_rate": 8.537634137001482e-08,
|
| 2445 |
+
"loss": 0.2761,
|
| 2446 |
+
"step": 1740
|
| 2447 |
+
},
|
| 2448 |
+
{
|
| 2449 |
+
"epoch": 2.9131886477462436,
|
| 2450 |
+
"grad_norm": 1.1730722188949585,
|
| 2451 |
+
"learning_rate": 7.130193869003765e-08,
|
| 2452 |
+
"loss": 0.2933,
|
| 2453 |
+
"step": 1745
|
| 2454 |
+
},
|
| 2455 |
+
{
|
| 2456 |
+
"epoch": 2.9215358931552586,
|
| 2457 |
+
"grad_norm": 1.29822838306427,
|
| 2458 |
+
"learning_rate": 5.849166555468799e-08,
|
| 2459 |
+
"loss": 0.3106,
|
| 2460 |
+
"step": 1750
|
| 2461 |
+
},
|
| 2462 |
+
{
|
| 2463 |
+
"epoch": 2.9298831385642736,
|
| 2464 |
+
"grad_norm": 1.1164788007736206,
|
| 2465 |
+
"learning_rate": 4.694660670991391e-08,
|
| 2466 |
+
"loss": 0.2957,
|
| 2467 |
+
"step": 1755
|
| 2468 |
+
},
|
| 2469 |
+
{
|
| 2470 |
+
"epoch": 2.9382303839732886,
|
| 2471 |
+
"grad_norm": 1.1868785619735718,
|
| 2472 |
+
"learning_rate": 3.666773976607973e-08,
|
| 2473 |
+
"loss": 0.2938,
|
| 2474 |
+
"step": 1760
|
| 2475 |
+
},
|
| 2476 |
+
{
|
| 2477 |
+
"epoch": 2.946577629382304,
|
| 2478 |
+
"grad_norm": 1.1611711978912354,
|
| 2479 |
+
"learning_rate": 2.7655935115188936e-08,
|
| 2480 |
+
"loss": 0.3024,
|
| 2481 |
+
"step": 1765
|
| 2482 |
+
},
|
| 2483 |
+
{
|
| 2484 |
+
"epoch": 2.9549248747913186,
|
| 2485 |
+
"grad_norm": 1.1683493852615356,
|
| 2486 |
+
"learning_rate": 1.9911955857176424e-08,
|
| 2487 |
+
"loss": 0.297,
|
| 2488 |
+
"step": 1770
|
| 2489 |
+
},
|
| 2490 |
+
{
|
| 2491 |
+
"epoch": 2.963272120200334,
|
| 2492 |
+
"grad_norm": 1.224556803703308,
|
| 2493 |
+
"learning_rate": 1.3436457735288566e-08,
|
| 2494 |
+
"loss": 0.2641,
|
| 2495 |
+
"step": 1775
|
| 2496 |
+
},
|
| 2497 |
+
{
|
| 2498 |
+
"epoch": 2.971619365609349,
|
| 2499 |
+
"grad_norm": 1.212898850440979,
|
| 2500 |
+
"learning_rate": 8.229989080565936e-09,
|
| 2501 |
+
"loss": 0.314,
|
| 2502 |
+
"step": 1780
|
| 2503 |
+
},
|
| 2504 |
+
{
|
| 2505 |
+
"epoch": 2.979966611018364,
|
| 2506 |
+
"grad_norm": 1.1555767059326172,
|
| 2507 |
+
"learning_rate": 4.292990765400462e-09,
|
| 2508 |
+
"loss": 0.2811,
|
| 2509 |
+
"step": 1785
|
| 2510 |
+
},
|
| 2511 |
+
{
|
| 2512 |
+
"epoch": 2.988313856427379,
|
| 2513 |
+
"grad_norm": 1.3482564687728882,
|
| 2514 |
+
"learning_rate": 1.6257961662102805e-09,
|
| 2515 |
+
"loss": 0.3173,
|
| 2516 |
+
"step": 1790
|
| 2517 |
+
},
|
| 2518 |
+
{
|
| 2519 |
+
"epoch": 2.996661101836394,
|
| 2520 |
+
"grad_norm": 1.1563059091567993,
|
| 2521 |
+
"learning_rate": 2.2863113520898982e-10,
|
| 2522 |
+
"loss": 0.2996,
|
| 2523 |
+
"step": 1795
|
| 2524 |
+
},
|
| 2525 |
+
{
|
| 2526 |
+
"epoch": 3.0,
|
| 2527 |
+
"step": 1797,
|
| 2528 |
+
"total_flos": 2.607849924611539e+18,
|
| 2529 |
+
"train_loss": 0.5861472732900047,
|
| 2530 |
+
"train_runtime": 1168.8941,
|
| 2531 |
+
"train_samples_per_second": 49.162,
|
| 2532 |
+
"train_steps_per_second": 1.537
|
| 2533 |
+
}
|
| 2534 |
+
],
|
| 2535 |
+
"logging_steps": 5,
|
| 2536 |
+
"max_steps": 1797,
|
| 2537 |
+
"num_input_tokens_seen": 0,
|
| 2538 |
+
"num_train_epochs": 3,
|
| 2539 |
+
"save_steps": 2000,
|
| 2540 |
+
"stateful_callbacks": {
|
| 2541 |
+
"TrainerControl": {
|
| 2542 |
+
"args": {
|
| 2543 |
+
"should_epoch_stop": false,
|
| 2544 |
+
"should_evaluate": false,
|
| 2545 |
+
"should_log": false,
|
| 2546 |
+
"should_save": false,
|
| 2547 |
+
"should_training_stop": false
|
| 2548 |
+
},
|
| 2549 |
+
"attributes": {}
|
| 2550 |
+
}
|
| 2551 |
+
},
|
| 2552 |
+
"total_flos": 2.607849924611539e+18,
|
| 2553 |
+
"train_batch_size": 2,
|
| 2554 |
+
"trial_name": null,
|
| 2555 |
+
"trial_params": null
|
| 2556 |
+
}
|
23_128_e3_3e-5/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:cd4fa71e2df715816d961d2febf7fe5fdc85f8722ebeca078ede9200ba9b6cc7
|
| 3 |
+
size 8145
|
23_128_e3_3e-5/vocab.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|