Upload 13 files
Browse files- README.md +52 -3
- adapter_config.json +31 -0
- adapter_model.safetensors +3 -0
- all_results.json +14 -0
- config.json +28 -0
- eval_results.json +8 -0
- special_tokens_map.json +24 -0
- tokenizer.json +0 -0
- tokenizer.model +3 -0
- tokenizer_config.json +42 -0
- train_results.json +9 -0
- trainer_state.json +2425 -0
- training_args.bin +3 -0
README.md
CHANGED
|
@@ -1,3 +1,52 @@
|
|
| 1 |
-
---
|
| 2 |
-
license: apache-2.0
|
| 3 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
library_name: peft
|
| 4 |
+
tags:
|
| 5 |
+
- alignment-handbook
|
| 6 |
+
- generated_from_trainer
|
| 7 |
+
base_model: h2oai/h2o-danube-1.8b-base
|
| 8 |
+
datasets:
|
| 9 |
+
- Ritvik19/open-hermes-2_5-reformatted
|
| 10 |
+
model-index:
|
| 11 |
+
- name: openhermes-danube-sft-qlora
|
| 12 |
+
results: []
|
| 13 |
+
---
|
| 14 |
+
|
| 15 |
+
**Note**: This model card has been generated automatically according to the information the Trainer had access to.
|
| 16 |
+
Visit the [model card](https://ritvik19.github.io/small-llms/) to see the full description
|
| 17 |
+
|
| 18 |
+
# openhermes-danube-sft-qlora
|
| 19 |
+
|
| 20 |
+
This model is a fine-tuned version of [h2oai/h2o-danube-1.8b-base](https://huggingface.co/h2oai/h2o-danube-1.8b-base) on the Ritvik19/open-hermes-2_5-reformatted dataset.
|
| 21 |
+
It achieves the following results on the evaluation set:
|
| 22 |
+
- Loss: 1.1863
|
| 23 |
+
|
| 24 |
+
### Training hyperparameters
|
| 25 |
+
|
| 26 |
+
The following hyperparameters were used during training:
|
| 27 |
+
- learning_rate: 0.0002
|
| 28 |
+
- train_batch_size: 1
|
| 29 |
+
- eval_batch_size: 1
|
| 30 |
+
- seed: 42
|
| 31 |
+
- distributed_type: multi-GPU
|
| 32 |
+
- gradient_accumulation_steps: 128
|
| 33 |
+
- total_train_batch_size: 128
|
| 34 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 35 |
+
- lr_scheduler_type: cosine
|
| 36 |
+
- lr_scheduler_warmup_ratio: 0.1
|
| 37 |
+
- num_epochs: 1
|
| 38 |
+
|
| 39 |
+
### Training results
|
| 40 |
+
|
| 41 |
+
| Training Loss | Epoch | Step | Validation Loss |
|
| 42 |
+
|:-------------:|:------:|:----:|:---------------:|
|
| 43 |
+
| 1.0838 | 0.9995 | 1704 | 1.1863 |
|
| 44 |
+
|
| 45 |
+
|
| 46 |
+
### Framework versions
|
| 47 |
+
|
| 48 |
+
- PEFT 0.7.1
|
| 49 |
+
- Transformers 4.40.1
|
| 50 |
+
- Pytorch 2.1.2+cu121
|
| 51 |
+
- Datasets 2.19.0
|
| 52 |
+
- Tokenizers 0.19.1
|
adapter_config.json
ADDED
|
@@ -0,0 +1,31 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"alpha_pattern": {},
|
| 3 |
+
"auto_mapping": null,
|
| 4 |
+
"base_model_name_or_path": "h2oai/h2o-danube-1.8b-base",
|
| 5 |
+
"bias": "none",
|
| 6 |
+
"fan_in_fan_out": false,
|
| 7 |
+
"inference_mode": true,
|
| 8 |
+
"init_lora_weights": true,
|
| 9 |
+
"layers_pattern": null,
|
| 10 |
+
"layers_to_transform": null,
|
| 11 |
+
"loftq_config": {},
|
| 12 |
+
"lora_alpha": 16,
|
| 13 |
+
"lora_dropout": 0.05,
|
| 14 |
+
"megatron_config": null,
|
| 15 |
+
"megatron_core": "megatron.core",
|
| 16 |
+
"modules_to_save": null,
|
| 17 |
+
"peft_type": "LORA",
|
| 18 |
+
"r": 16,
|
| 19 |
+
"rank_pattern": {},
|
| 20 |
+
"revision": null,
|
| 21 |
+
"target_modules": [
|
| 22 |
+
"gate_proj",
|
| 23 |
+
"v_proj",
|
| 24 |
+
"q_proj",
|
| 25 |
+
"k_proj",
|
| 26 |
+
"o_proj",
|
| 27 |
+
"down_proj",
|
| 28 |
+
"up_proj"
|
| 29 |
+
],
|
| 30 |
+
"task_type": "CAUSAL_LM"
|
| 31 |
+
}
|
adapter_model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:75a8dd3562000e4d3ca36546916dfee163a553fa5648780c990a812a385d5641
|
| 3 |
+
size 34647816
|
all_results.json
ADDED
|
@@ -0,0 +1,14 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 0.9994547062516325,
|
| 3 |
+
"eval_loss": 1.1863435506820679,
|
| 4 |
+
"eval_runtime": 1962.7404,
|
| 5 |
+
"eval_samples": 23109,
|
| 6 |
+
"eval_samples_per_second": 8.229,
|
| 7 |
+
"eval_steps_per_second": 8.229,
|
| 8 |
+
"total_flos": 4.734725057610252e+18,
|
| 9 |
+
"train_loss": 0.12850537406446788,
|
| 10 |
+
"train_runtime": 13780.2764,
|
| 11 |
+
"train_samples": 1001538,
|
| 12 |
+
"train_samples_per_second": 15.836,
|
| 13 |
+
"train_steps_per_second": 0.124
|
| 14 |
+
}
|
config.json
ADDED
|
@@ -0,0 +1,28 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "h2oai/h2o-danube-1.8b-base",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"MistralForCausalLM"
|
| 5 |
+
],
|
| 6 |
+
"attention_dropout": 0.0,
|
| 7 |
+
"bos_token_id": 1,
|
| 8 |
+
"eos_token_id": 2,
|
| 9 |
+
"hidden_act": "silu",
|
| 10 |
+
"hidden_size": 2560,
|
| 11 |
+
"initializer_range": 0.02,
|
| 12 |
+
"intermediate_size": 6912,
|
| 13 |
+
"max_position_embeddings": 16384,
|
| 14 |
+
"model_type": "mistral",
|
| 15 |
+
"num_attention_heads": 32,
|
| 16 |
+
"num_hidden_layers": 24,
|
| 17 |
+
"num_key_value_heads": 8,
|
| 18 |
+
"pretraining_tp": 1,
|
| 19 |
+
"rms_norm_eps": 1e-05,
|
| 20 |
+
"rope_scaling": null,
|
| 21 |
+
"rope_theta": 10000.0,
|
| 22 |
+
"sliding_window": 4096,
|
| 23 |
+
"tie_word_embeddings": false,
|
| 24 |
+
"torch_dtype": "bfloat16",
|
| 25 |
+
"transformers_version": "4.40.1",
|
| 26 |
+
"use_cache": true,
|
| 27 |
+
"vocab_size": 32000
|
| 28 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 0.9994547062516325,
|
| 3 |
+
"eval_loss": 1.1863435506820679,
|
| 4 |
+
"eval_runtime": 1962.7404,
|
| 5 |
+
"eval_samples": 23109,
|
| 6 |
+
"eval_samples_per_second": 8.229,
|
| 7 |
+
"eval_steps_per_second": 8.229
|
| 8 |
+
}
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1,24 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"content": "<s>",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": false,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"eos_token": {
|
| 10 |
+
"content": "</s>",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"pad_token": "</s>",
|
| 17 |
+
"unk_token": {
|
| 18 |
+
"content": "<unk>",
|
| 19 |
+
"lstrip": false,
|
| 20 |
+
"normalized": false,
|
| 21 |
+
"rstrip": false,
|
| 22 |
+
"single_word": false
|
| 23 |
+
}
|
| 24 |
+
}
|
tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tokenizer.model
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9e556afd44213b6bd1be2b850ebbbd98f5481437a8021afaf58ee7fb1818d347
|
| 3 |
+
size 499723
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,42 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_bos_token": false,
|
| 3 |
+
"add_eos_token": false,
|
| 4 |
+
"added_tokens_decoder": {
|
| 5 |
+
"0": {
|
| 6 |
+
"content": "<unk>",
|
| 7 |
+
"lstrip": false,
|
| 8 |
+
"normalized": false,
|
| 9 |
+
"rstrip": false,
|
| 10 |
+
"single_word": false,
|
| 11 |
+
"special": true
|
| 12 |
+
},
|
| 13 |
+
"1": {
|
| 14 |
+
"content": "<s>",
|
| 15 |
+
"lstrip": false,
|
| 16 |
+
"normalized": false,
|
| 17 |
+
"rstrip": false,
|
| 18 |
+
"single_word": false,
|
| 19 |
+
"special": true
|
| 20 |
+
},
|
| 21 |
+
"2": {
|
| 22 |
+
"content": "</s>",
|
| 23 |
+
"lstrip": false,
|
| 24 |
+
"normalized": false,
|
| 25 |
+
"rstrip": false,
|
| 26 |
+
"single_word": false,
|
| 27 |
+
"special": true
|
| 28 |
+
}
|
| 29 |
+
},
|
| 30 |
+
"bos_token": "<s>",
|
| 31 |
+
"chat_template": "{% for message in messages %}\n{% if message['role'] == 'user' %}\n{{ '<|user|>\n' + message['content'] + eos_token }}\n{% elif message['role'] == 'system' %}\n{{ '<|system|>\n' + message['content'] + eos_token }}\n{% elif message['role'] == 'assistant' %}\n{{ '<|assistant|>\n' + message['content'] + eos_token }}\n{% endif %}\n{% if loop.last and add_generation_prompt %}\n{{ '<|assistant|>' }}\n{% endif %}\n{% endfor %}",
|
| 32 |
+
"clean_up_tokenization_spaces": false,
|
| 33 |
+
"eos_token": "</s>",
|
| 34 |
+
"legacy": false,
|
| 35 |
+
"model_max_length": 2048,
|
| 36 |
+
"pad_token": "</s>",
|
| 37 |
+
"padding_side": "left",
|
| 38 |
+
"sp_model_kwargs": {},
|
| 39 |
+
"tokenizer_class": "LlamaTokenizer",
|
| 40 |
+
"unk_token": "<unk>",
|
| 41 |
+
"use_default_system_prompt": false
|
| 42 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 0.9994547062516325,
|
| 3 |
+
"total_flos": 4.734725057610252e+18,
|
| 4 |
+
"train_loss": 0.12850537406446788,
|
| 5 |
+
"train_runtime": 13780.2764,
|
| 6 |
+
"train_samples": 1001538,
|
| 7 |
+
"train_samples_per_second": 15.836,
|
| 8 |
+
"train_steps_per_second": 0.124
|
| 9 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,2425 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.9994547062516325,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 1704,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.0005865344520256058,
|
| 13 |
+
"grad_norm": 0.134765625,
|
| 14 |
+
"learning_rate": 1.1695906432748538e-06,
|
| 15 |
+
"loss": 1.5828,
|
| 16 |
+
"step": 1
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.0029326722601280297,
|
| 20 |
+
"grad_norm": 0.1298828125,
|
| 21 |
+
"learning_rate": 5.8479532163742686e-06,
|
| 22 |
+
"loss": 1.5339,
|
| 23 |
+
"step": 5
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.005865344520256059,
|
| 27 |
+
"grad_norm": 0.12890625,
|
| 28 |
+
"learning_rate": 1.1695906432748537e-05,
|
| 29 |
+
"loss": 1.5449,
|
| 30 |
+
"step": 10
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.008798016780384088,
|
| 34 |
+
"grad_norm": 0.1328125,
|
| 35 |
+
"learning_rate": 1.7543859649122806e-05,
|
| 36 |
+
"loss": 1.5426,
|
| 37 |
+
"step": 15
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.011730689040512119,
|
| 41 |
+
"grad_norm": 0.1279296875,
|
| 42 |
+
"learning_rate": 2.3391812865497074e-05,
|
| 43 |
+
"loss": 1.5446,
|
| 44 |
+
"step": 20
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.014663361300640148,
|
| 48 |
+
"grad_norm": 0.142578125,
|
| 49 |
+
"learning_rate": 2.9239766081871346e-05,
|
| 50 |
+
"loss": 1.5188,
|
| 51 |
+
"step": 25
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.017596033560768175,
|
| 55 |
+
"grad_norm": 0.1298828125,
|
| 56 |
+
"learning_rate": 3.508771929824561e-05,
|
| 57 |
+
"loss": 1.5175,
|
| 58 |
+
"step": 30
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.020528705820896206,
|
| 62 |
+
"grad_norm": 0.119140625,
|
| 63 |
+
"learning_rate": 4.093567251461988e-05,
|
| 64 |
+
"loss": 1.4769,
|
| 65 |
+
"step": 35
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.023461378081024237,
|
| 69 |
+
"grad_norm": 0.1171875,
|
| 70 |
+
"learning_rate": 4.678362573099415e-05,
|
| 71 |
+
"loss": 1.4503,
|
| 72 |
+
"step": 40
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.026394050341152265,
|
| 76 |
+
"grad_norm": 0.0947265625,
|
| 77 |
+
"learning_rate": 5.2631578947368424e-05,
|
| 78 |
+
"loss": 1.4078,
|
| 79 |
+
"step": 45
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.029326722601280296,
|
| 83 |
+
"grad_norm": 0.08642578125,
|
| 84 |
+
"learning_rate": 5.847953216374269e-05,
|
| 85 |
+
"loss": 1.3803,
|
| 86 |
+
"step": 50
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.03225939486140832,
|
| 90 |
+
"grad_norm": 0.07373046875,
|
| 91 |
+
"learning_rate": 6.432748538011695e-05,
|
| 92 |
+
"loss": 1.365,
|
| 93 |
+
"step": 55
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.03519206712153635,
|
| 97 |
+
"grad_norm": 0.06689453125,
|
| 98 |
+
"learning_rate": 7.017543859649122e-05,
|
| 99 |
+
"loss": 1.3275,
|
| 100 |
+
"step": 60
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.038124739381664385,
|
| 104 |
+
"grad_norm": 0.0634765625,
|
| 105 |
+
"learning_rate": 7.602339181286549e-05,
|
| 106 |
+
"loss": 1.3026,
|
| 107 |
+
"step": 65
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.04105741164179241,
|
| 111 |
+
"grad_norm": 0.06005859375,
|
| 112 |
+
"learning_rate": 8.187134502923976e-05,
|
| 113 |
+
"loss": 1.2994,
|
| 114 |
+
"step": 70
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.04399008390192044,
|
| 118 |
+
"grad_norm": 0.0634765625,
|
| 119 |
+
"learning_rate": 8.771929824561403e-05,
|
| 120 |
+
"loss": 1.2961,
|
| 121 |
+
"step": 75
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.046922756162048475,
|
| 125 |
+
"grad_norm": 0.06201171875,
|
| 126 |
+
"learning_rate": 9.35672514619883e-05,
|
| 127 |
+
"loss": 1.2742,
|
| 128 |
+
"step": 80
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.0498554284221765,
|
| 132 |
+
"grad_norm": 0.04833984375,
|
| 133 |
+
"learning_rate": 9.941520467836257e-05,
|
| 134 |
+
"loss": 1.2694,
|
| 135 |
+
"step": 85
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.05278810068230453,
|
| 139 |
+
"grad_norm": 0.0537109375,
|
| 140 |
+
"learning_rate": 0.00010526315789473685,
|
| 141 |
+
"loss": 1.2601,
|
| 142 |
+
"step": 90
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.05572077294243256,
|
| 146 |
+
"grad_norm": 0.05419921875,
|
| 147 |
+
"learning_rate": 0.00011111111111111112,
|
| 148 |
+
"loss": 1.2419,
|
| 149 |
+
"step": 95
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.05865344520256059,
|
| 153 |
+
"grad_norm": 0.053466796875,
|
| 154 |
+
"learning_rate": 0.00011695906432748539,
|
| 155 |
+
"loss": 1.2357,
|
| 156 |
+
"step": 100
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.06158611746268862,
|
| 160 |
+
"grad_norm": 0.060302734375,
|
| 161 |
+
"learning_rate": 0.00012280701754385965,
|
| 162 |
+
"loss": 1.2227,
|
| 163 |
+
"step": 105
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.06451878972281665,
|
| 167 |
+
"grad_norm": 0.052978515625,
|
| 168 |
+
"learning_rate": 0.0001286549707602339,
|
| 169 |
+
"loss": 1.2327,
|
| 170 |
+
"step": 110
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.06745146198294467,
|
| 174 |
+
"grad_norm": 0.061767578125,
|
| 175 |
+
"learning_rate": 0.0001345029239766082,
|
| 176 |
+
"loss": 1.2221,
|
| 177 |
+
"step": 115
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.0703841342430727,
|
| 181 |
+
"grad_norm": 0.05859375,
|
| 182 |
+
"learning_rate": 0.00014035087719298245,
|
| 183 |
+
"loss": 1.209,
|
| 184 |
+
"step": 120
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 0.07331680650320074,
|
| 188 |
+
"grad_norm": 0.064453125,
|
| 189 |
+
"learning_rate": 0.00014619883040935673,
|
| 190 |
+
"loss": 1.2159,
|
| 191 |
+
"step": 125
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.07624947876332877,
|
| 195 |
+
"grad_norm": 0.0791015625,
|
| 196 |
+
"learning_rate": 0.00015204678362573098,
|
| 197 |
+
"loss": 1.1813,
|
| 198 |
+
"step": 130
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 0.0791821510234568,
|
| 202 |
+
"grad_norm": 0.0693359375,
|
| 203 |
+
"learning_rate": 0.00015789473684210527,
|
| 204 |
+
"loss": 1.2059,
|
| 205 |
+
"step": 135
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.08211482328358483,
|
| 209 |
+
"grad_norm": 0.06396484375,
|
| 210 |
+
"learning_rate": 0.00016374269005847952,
|
| 211 |
+
"loss": 1.2146,
|
| 212 |
+
"step": 140
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.08504749554371285,
|
| 216 |
+
"grad_norm": 0.09326171875,
|
| 217 |
+
"learning_rate": 0.0001695906432748538,
|
| 218 |
+
"loss": 1.1965,
|
| 219 |
+
"step": 145
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.08798016780384088,
|
| 223 |
+
"grad_norm": 0.0791015625,
|
| 224 |
+
"learning_rate": 0.00017543859649122806,
|
| 225 |
+
"loss": 1.1822,
|
| 226 |
+
"step": 150
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 0.09091284006396891,
|
| 230 |
+
"grad_norm": 0.08349609375,
|
| 231 |
+
"learning_rate": 0.00018128654970760234,
|
| 232 |
+
"loss": 1.1824,
|
| 233 |
+
"step": 155
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.09384551232409695,
|
| 237 |
+
"grad_norm": 0.078125,
|
| 238 |
+
"learning_rate": 0.0001871345029239766,
|
| 239 |
+
"loss": 1.1992,
|
| 240 |
+
"step": 160
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 0.09677818458422498,
|
| 244 |
+
"grad_norm": 0.0791015625,
|
| 245 |
+
"learning_rate": 0.00019298245614035088,
|
| 246 |
+
"loss": 1.1806,
|
| 247 |
+
"step": 165
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.099710856844353,
|
| 251 |
+
"grad_norm": 0.07666015625,
|
| 252 |
+
"learning_rate": 0.00019883040935672513,
|
| 253 |
+
"loss": 1.1926,
|
| 254 |
+
"step": 170
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.10264352910448103,
|
| 258 |
+
"grad_norm": 0.07666015625,
|
| 259 |
+
"learning_rate": 0.00019999664028072614,
|
| 260 |
+
"loss": 1.1773,
|
| 261 |
+
"step": 175
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.10557620136460906,
|
| 265 |
+
"grad_norm": 0.0830078125,
|
| 266 |
+
"learning_rate": 0.000199982991808088,
|
| 267 |
+
"loss": 1.1422,
|
| 268 |
+
"step": 180
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 0.10850887362473709,
|
| 272 |
+
"grad_norm": 0.11279296875,
|
| 273 |
+
"learning_rate": 0.00019995884603149402,
|
| 274 |
+
"loss": 1.1739,
|
| 275 |
+
"step": 185
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.11144154588486511,
|
| 279 |
+
"grad_norm": 0.0859375,
|
| 280 |
+
"learning_rate": 0.00019992420548603092,
|
| 281 |
+
"loss": 1.1652,
|
| 282 |
+
"step": 190
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.11437421814499316,
|
| 286 |
+
"grad_norm": 0.08154296875,
|
| 287 |
+
"learning_rate": 0.00019987907380864062,
|
| 288 |
+
"loss": 1.1597,
|
| 289 |
+
"step": 195
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.11730689040512118,
|
| 293 |
+
"grad_norm": 0.07861328125,
|
| 294 |
+
"learning_rate": 0.00019982345573773844,
|
| 295 |
+
"loss": 1.1497,
|
| 296 |
+
"step": 200
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.12023956266524921,
|
| 300 |
+
"grad_norm": 0.08544921875,
|
| 301 |
+
"learning_rate": 0.00019975735711271552,
|
| 302 |
+
"loss": 1.1419,
|
| 303 |
+
"step": 205
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.12317223492537724,
|
| 307 |
+
"grad_norm": 0.080078125,
|
| 308 |
+
"learning_rate": 0.00019968078487332566,
|
| 309 |
+
"loss": 1.1583,
|
| 310 |
+
"step": 210
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 0.12610490718550527,
|
| 314 |
+
"grad_norm": 0.08203125,
|
| 315 |
+
"learning_rate": 0.000199593747058957,
|
| 316 |
+
"loss": 1.1501,
|
| 317 |
+
"step": 215
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.1290375794456333,
|
| 321 |
+
"grad_norm": 0.0830078125,
|
| 322 |
+
"learning_rate": 0.00019949625280778777,
|
| 323 |
+
"loss": 1.1552,
|
| 324 |
+
"step": 220
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 0.13197025170576132,
|
| 328 |
+
"grad_norm": 0.080078125,
|
| 329 |
+
"learning_rate": 0.00019938831235582672,
|
| 330 |
+
"loss": 1.1285,
|
| 331 |
+
"step": 225
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.13490292396588935,
|
| 335 |
+
"grad_norm": 0.07861328125,
|
| 336 |
+
"learning_rate": 0.0001992699370358387,
|
| 337 |
+
"loss": 1.1525,
|
| 338 |
+
"step": 230
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.13783559622601738,
|
| 342 |
+
"grad_norm": 0.0830078125,
|
| 343 |
+
"learning_rate": 0.00019914113927615472,
|
| 344 |
+
"loss": 1.1405,
|
| 345 |
+
"step": 235
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.1407682684861454,
|
| 349 |
+
"grad_norm": 0.07861328125,
|
| 350 |
+
"learning_rate": 0.00019900193259936704,
|
| 351 |
+
"loss": 1.1432,
|
| 352 |
+
"step": 240
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 0.14370094074627343,
|
| 356 |
+
"grad_norm": 0.08056640625,
|
| 357 |
+
"learning_rate": 0.00019885233162090946,
|
| 358 |
+
"loss": 1.1523,
|
| 359 |
+
"step": 245
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.1466336130064015,
|
| 363 |
+
"grad_norm": 0.0751953125,
|
| 364 |
+
"learning_rate": 0.00019869235204752285,
|
| 365 |
+
"loss": 1.1592,
|
| 366 |
+
"step": 250
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 0.14956628526652951,
|
| 370 |
+
"grad_norm": 0.08544921875,
|
| 371 |
+
"learning_rate": 0.00019852201067560606,
|
| 372 |
+
"loss": 1.1535,
|
| 373 |
+
"step": 255
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.15249895752665754,
|
| 377 |
+
"grad_norm": 0.0771484375,
|
| 378 |
+
"learning_rate": 0.00019834132538945246,
|
| 379 |
+
"loss": 1.1429,
|
| 380 |
+
"step": 260
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.15543162978678557,
|
| 384 |
+
"grad_norm": 0.07763671875,
|
| 385 |
+
"learning_rate": 0.00019815031515937225,
|
| 386 |
+
"loss": 1.1385,
|
| 387 |
+
"step": 265
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.1583643020469136,
|
| 391 |
+
"grad_norm": 0.08642578125,
|
| 392 |
+
"learning_rate": 0.00019794900003970077,
|
| 393 |
+
"loss": 1.1356,
|
| 394 |
+
"step": 270
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 0.16129697430704162,
|
| 398 |
+
"grad_norm": 0.0849609375,
|
| 399 |
+
"learning_rate": 0.00019773740116669288,
|
| 400 |
+
"loss": 1.1311,
|
| 401 |
+
"step": 275
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.16422964656716965,
|
| 405 |
+
"grad_norm": 0.0869140625,
|
| 406 |
+
"learning_rate": 0.00019751554075630404,
|
| 407 |
+
"loss": 1.1243,
|
| 408 |
+
"step": 280
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 0.16716231882729768,
|
| 412 |
+
"grad_norm": 0.0908203125,
|
| 413 |
+
"learning_rate": 0.0001972834421018576,
|
| 414 |
+
"loss": 1.1475,
|
| 415 |
+
"step": 285
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 0.1700949910874257,
|
| 419 |
+
"grad_norm": 0.08056640625,
|
| 420 |
+
"learning_rate": 0.0001970411295715994,
|
| 421 |
+
"loss": 1.1338,
|
| 422 |
+
"step": 290
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.17302766334755373,
|
| 426 |
+
"grad_norm": 0.0810546875,
|
| 427 |
+
"learning_rate": 0.0001967886286061393,
|
| 428 |
+
"loss": 1.1238,
|
| 429 |
+
"step": 295
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.17596033560768176,
|
| 433 |
+
"grad_norm": 0.08642578125,
|
| 434 |
+
"learning_rate": 0.00019652596571578004,
|
| 435 |
+
"loss": 1.1322,
|
| 436 |
+
"step": 300
|
| 437 |
+
},
|
| 438 |
+
{
|
| 439 |
+
"epoch": 0.1788930078678098,
|
| 440 |
+
"grad_norm": 0.0712890625,
|
| 441 |
+
"learning_rate": 0.00019625316847773395,
|
| 442 |
+
"loss": 1.1305,
|
| 443 |
+
"step": 305
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.18182568012793782,
|
| 447 |
+
"grad_norm": 0.07666015625,
|
| 448 |
+
"learning_rate": 0.0001959702655332277,
|
| 449 |
+
"loss": 1.1273,
|
| 450 |
+
"step": 310
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 0.18475835238806584,
|
| 454 |
+
"grad_norm": 0.08349609375,
|
| 455 |
+
"learning_rate": 0.00019567728658449504,
|
| 456 |
+
"loss": 1.1306,
|
| 457 |
+
"step": 315
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 0.1876910246481939,
|
| 461 |
+
"grad_norm": 0.0830078125,
|
| 462 |
+
"learning_rate": 0.00019537426239165853,
|
| 463 |
+
"loss": 1.1239,
|
| 464 |
+
"step": 320
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.19062369690832193,
|
| 468 |
+
"grad_norm": 0.0869140625,
|
| 469 |
+
"learning_rate": 0.00019506122476949981,
|
| 470 |
+
"loss": 1.1234,
|
| 471 |
+
"step": 325
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.19355636916844995,
|
| 475 |
+
"grad_norm": 0.09228515625,
|
| 476 |
+
"learning_rate": 0.00019473820658411957,
|
| 477 |
+
"loss": 1.1115,
|
| 478 |
+
"step": 330
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 0.19648904142857798,
|
| 482 |
+
"grad_norm": 0.09033203125,
|
| 483 |
+
"learning_rate": 0.0001944052417494867,
|
| 484 |
+
"loss": 1.1127,
|
| 485 |
+
"step": 335
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 0.199421713688706,
|
| 489 |
+
"grad_norm": 0.08740234375,
|
| 490 |
+
"learning_rate": 0.0001940623652238777,
|
| 491 |
+
"loss": 1.1315,
|
| 492 |
+
"step": 340
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 0.20235438594883404,
|
| 496 |
+
"grad_norm": 0.0771484375,
|
| 497 |
+
"learning_rate": 0.00019370961300620637,
|
| 498 |
+
"loss": 1.1211,
|
| 499 |
+
"step": 345
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 0.20528705820896206,
|
| 503 |
+
"grad_norm": 0.07421875,
|
| 504 |
+
"learning_rate": 0.00019334702213224446,
|
| 505 |
+
"loss": 1.1162,
|
| 506 |
+
"step": 350
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 0.2082197304690901,
|
| 510 |
+
"grad_norm": 0.07421875,
|
| 511 |
+
"learning_rate": 0.00019297463067073287,
|
| 512 |
+
"loss": 1.129,
|
| 513 |
+
"step": 355
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.21115240272921812,
|
| 517 |
+
"grad_norm": 0.0791015625,
|
| 518 |
+
"learning_rate": 0.000192592477719385,
|
| 519 |
+
"loss": 1.108,
|
| 520 |
+
"step": 360
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 0.21408507498934615,
|
| 524 |
+
"grad_norm": 0.080078125,
|
| 525 |
+
"learning_rate": 0.00019220060340078188,
|
| 526 |
+
"loss": 1.1167,
|
| 527 |
+
"step": 365
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 0.21701774724947417,
|
| 531 |
+
"grad_norm": 0.0869140625,
|
| 532 |
+
"learning_rate": 0.00019179904885815958,
|
| 533 |
+
"loss": 1.1485,
|
| 534 |
+
"step": 370
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 0.2199504195096022,
|
| 538 |
+
"grad_norm": 0.0751953125,
|
| 539 |
+
"learning_rate": 0.00019138785625108957,
|
| 540 |
+
"loss": 1.1183,
|
| 541 |
+
"step": 375
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 0.22288309176973023,
|
| 545 |
+
"grad_norm": 0.0869140625,
|
| 546 |
+
"learning_rate": 0.00019096706875105235,
|
| 547 |
+
"loss": 1.1176,
|
| 548 |
+
"step": 380
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 0.22581576402985826,
|
| 552 |
+
"grad_norm": 0.08203125,
|
| 553 |
+
"learning_rate": 0.0001905367305369048,
|
| 554 |
+
"loss": 1.1184,
|
| 555 |
+
"step": 385
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.2287484362899863,
|
| 559 |
+
"grad_norm": 0.0732421875,
|
| 560 |
+
"learning_rate": 0.0001900968867902419,
|
| 561 |
+
"loss": 1.1197,
|
| 562 |
+
"step": 390
|
| 563 |
+
},
|
| 564 |
+
{
|
| 565 |
+
"epoch": 0.23168110855011434,
|
| 566 |
+
"grad_norm": 0.09521484375,
|
| 567 |
+
"learning_rate": 0.000189647583690653,
|
| 568 |
+
"loss": 1.1265,
|
| 569 |
+
"step": 395
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 0.23461378081024237,
|
| 573 |
+
"grad_norm": 0.08251953125,
|
| 574 |
+
"learning_rate": 0.00018918886841087334,
|
| 575 |
+
"loss": 1.1209,
|
| 576 |
+
"step": 400
|
| 577 |
+
},
|
| 578 |
+
{
|
| 579 |
+
"epoch": 0.2375464530703704,
|
| 580 |
+
"grad_norm": 0.0732421875,
|
| 581 |
+
"learning_rate": 0.00018872078911183146,
|
| 582 |
+
"loss": 1.1084,
|
| 583 |
+
"step": 405
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 0.24047912533049842,
|
| 587 |
+
"grad_norm": 0.08447265625,
|
| 588 |
+
"learning_rate": 0.00018824339493759263,
|
| 589 |
+
"loss": 1.132,
|
| 590 |
+
"step": 410
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 0.24341179759062645,
|
| 594 |
+
"grad_norm": 0.07666015625,
|
| 595 |
+
"learning_rate": 0.00018775673601019923,
|
| 596 |
+
"loss": 1.1122,
|
| 597 |
+
"step": 415
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 0.24634446985075448,
|
| 601 |
+
"grad_norm": 0.07666015625,
|
| 602 |
+
"learning_rate": 0.00018726086342440846,
|
| 603 |
+
"loss": 1.1161,
|
| 604 |
+
"step": 420
|
| 605 |
+
},
|
| 606 |
+
{
|
| 607 |
+
"epoch": 0.2492771421108825,
|
| 608 |
+
"grad_norm": 0.07275390625,
|
| 609 |
+
"learning_rate": 0.00018675582924232762,
|
| 610 |
+
"loss": 1.1183,
|
| 611 |
+
"step": 425
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 0.25220981437101053,
|
| 615 |
+
"grad_norm": 0.078125,
|
| 616 |
+
"learning_rate": 0.00018624168648794832,
|
| 617 |
+
"loss": 1.1029,
|
| 618 |
+
"step": 430
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"epoch": 0.2551424866311386,
|
| 622 |
+
"grad_norm": 0.07666015625,
|
| 623 |
+
"learning_rate": 0.0001857184891415794,
|
| 624 |
+
"loss": 1.1236,
|
| 625 |
+
"step": 435
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 0.2580751588912666,
|
| 629 |
+
"grad_norm": 0.0810546875,
|
| 630 |
+
"learning_rate": 0.00018518629213417929,
|
| 631 |
+
"loss": 1.1036,
|
| 632 |
+
"step": 440
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 0.26100783115139464,
|
| 636 |
+
"grad_norm": 0.07275390625,
|
| 637 |
+
"learning_rate": 0.00018464515134158896,
|
| 638 |
+
"loss": 1.1047,
|
| 639 |
+
"step": 445
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.26394050341152264,
|
| 643 |
+
"grad_norm": 0.07470703125,
|
| 644 |
+
"learning_rate": 0.00018409512357866548,
|
| 645 |
+
"loss": 1.1153,
|
| 646 |
+
"step": 450
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"epoch": 0.2668731756716507,
|
| 650 |
+
"grad_norm": 0.07666015625,
|
| 651 |
+
"learning_rate": 0.00018353626659331683,
|
| 652 |
+
"loss": 1.1238,
|
| 653 |
+
"step": 455
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 0.2698058479317787,
|
| 657 |
+
"grad_norm": 0.07568359375,
|
| 658 |
+
"learning_rate": 0.00018296863906043894,
|
| 659 |
+
"loss": 1.1149,
|
| 660 |
+
"step": 460
|
| 661 |
+
},
|
| 662 |
+
{
|
| 663 |
+
"epoch": 0.27273852019190675,
|
| 664 |
+
"grad_norm": 0.08056640625,
|
| 665 |
+
"learning_rate": 0.00018239230057575542,
|
| 666 |
+
"loss": 1.1112,
|
| 667 |
+
"step": 465
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 0.27567119245203475,
|
| 671 |
+
"grad_norm": 0.0751953125,
|
| 672 |
+
"learning_rate": 0.0001818073116495606,
|
| 673 |
+
"loss": 1.1043,
|
| 674 |
+
"step": 470
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 0.2786038647121628,
|
| 678 |
+
"grad_norm": 0.0732421875,
|
| 679 |
+
"learning_rate": 0.0001812137337003663,
|
| 680 |
+
"loss": 1.1055,
|
| 681 |
+
"step": 475
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 0.2815365369722908,
|
| 685 |
+
"grad_norm": 0.0732421875,
|
| 686 |
+
"learning_rate": 0.00018061162904845358,
|
| 687 |
+
"loss": 1.1059,
|
| 688 |
+
"step": 480
|
| 689 |
+
},
|
| 690 |
+
{
|
| 691 |
+
"epoch": 0.28446920923241886,
|
| 692 |
+
"grad_norm": 0.07666015625,
|
| 693 |
+
"learning_rate": 0.0001800010609093298,
|
| 694 |
+
"loss": 1.1062,
|
| 695 |
+
"step": 485
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"epoch": 0.28740188149254686,
|
| 699 |
+
"grad_norm": 0.07763671875,
|
| 700 |
+
"learning_rate": 0.00017938209338709123,
|
| 701 |
+
"loss": 1.115,
|
| 702 |
+
"step": 490
|
| 703 |
+
},
|
| 704 |
+
{
|
| 705 |
+
"epoch": 0.2903345537526749,
|
| 706 |
+
"grad_norm": 0.076171875,
|
| 707 |
+
"learning_rate": 0.00017875479146769305,
|
| 708 |
+
"loss": 1.0987,
|
| 709 |
+
"step": 495
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"epoch": 0.293267226012803,
|
| 713 |
+
"grad_norm": 0.08056640625,
|
| 714 |
+
"learning_rate": 0.0001781192210121262,
|
| 715 |
+
"loss": 1.1072,
|
| 716 |
+
"step": 500
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 0.296199898272931,
|
| 720 |
+
"grad_norm": 0.07421875,
|
| 721 |
+
"learning_rate": 0.00017747544874950272,
|
| 722 |
+
"loss": 1.1027,
|
| 723 |
+
"step": 505
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 0.29913257053305903,
|
| 727 |
+
"grad_norm": 0.08837890625,
|
| 728 |
+
"learning_rate": 0.00017682354227004963,
|
| 729 |
+
"loss": 1.1004,
|
| 730 |
+
"step": 510
|
| 731 |
+
},
|
| 732 |
+
{
|
| 733 |
+
"epoch": 0.302065242793187,
|
| 734 |
+
"grad_norm": 0.07763671875,
|
| 735 |
+
"learning_rate": 0.0001761635700180127,
|
| 736 |
+
"loss": 1.1012,
|
| 737 |
+
"step": 515
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 0.3049979150533151,
|
| 741 |
+
"grad_norm": 0.07861328125,
|
| 742 |
+
"learning_rate": 0.00017549560128447047,
|
| 743 |
+
"loss": 1.0992,
|
| 744 |
+
"step": 520
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 0.3079305873134431,
|
| 748 |
+
"grad_norm": 0.10888671875,
|
| 749 |
+
"learning_rate": 0.00017481970620005912,
|
| 750 |
+
"loss": 1.1039,
|
| 751 |
+
"step": 525
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 0.31086325957357114,
|
| 755 |
+
"grad_norm": 0.07421875,
|
| 756 |
+
"learning_rate": 0.00017413595572760961,
|
| 757 |
+
"loss": 1.1154,
|
| 758 |
+
"step": 530
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 0.31379593183369914,
|
| 762 |
+
"grad_norm": 0.07861328125,
|
| 763 |
+
"learning_rate": 0.00017344442165469714,
|
| 764 |
+
"loss": 1.0995,
|
| 765 |
+
"step": 535
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 0.3167286040938272,
|
| 769 |
+
"grad_norm": 0.07861328125,
|
| 770 |
+
"learning_rate": 0.00017274517658610398,
|
| 771 |
+
"loss": 1.0934,
|
| 772 |
+
"step": 540
|
| 773 |
+
},
|
| 774 |
+
{
|
| 775 |
+
"epoch": 0.3196612763539552,
|
| 776 |
+
"grad_norm": 0.08154296875,
|
| 777 |
+
"learning_rate": 0.0001720382939361969,
|
| 778 |
+
"loss": 1.0801,
|
| 779 |
+
"step": 545
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 0.32259394861408325,
|
| 783 |
+
"grad_norm": 0.07763671875,
|
| 784 |
+
"learning_rate": 0.00017132384792121905,
|
| 785 |
+
"loss": 1.0932,
|
| 786 |
+
"step": 550
|
| 787 |
+
},
|
| 788 |
+
{
|
| 789 |
+
"epoch": 0.32552662087421125,
|
| 790 |
+
"grad_norm": 0.076171875,
|
| 791 |
+
"learning_rate": 0.0001706019135514982,
|
| 792 |
+
"loss": 1.12,
|
| 793 |
+
"step": 555
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 0.3284592931343393,
|
| 797 |
+
"grad_norm": 0.0927734375,
|
| 798 |
+
"learning_rate": 0.00016987256662357106,
|
| 799 |
+
"loss": 1.129,
|
| 800 |
+
"step": 560
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 0.33139196539446736,
|
| 804 |
+
"grad_norm": 0.0869140625,
|
| 805 |
+
"learning_rate": 0.00016913588371222557,
|
| 806 |
+
"loss": 1.1059,
|
| 807 |
+
"step": 565
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 0.33432463765459536,
|
| 811 |
+
"grad_norm": 0.08642578125,
|
| 812 |
+
"learning_rate": 0.00016839194216246108,
|
| 813 |
+
"loss": 1.0923,
|
| 814 |
+
"step": 570
|
| 815 |
+
},
|
| 816 |
+
{
|
| 817 |
+
"epoch": 0.3372573099147234,
|
| 818 |
+
"grad_norm": 0.0849609375,
|
| 819 |
+
"learning_rate": 0.00016764082008136795,
|
| 820 |
+
"loss": 1.0909,
|
| 821 |
+
"step": 575
|
| 822 |
+
},
|
| 823 |
+
{
|
| 824 |
+
"epoch": 0.3401899821748514,
|
| 825 |
+
"grad_norm": 0.0771484375,
|
| 826 |
+
"learning_rate": 0.00016688259632992693,
|
| 827 |
+
"loss": 1.096,
|
| 828 |
+
"step": 580
|
| 829 |
+
},
|
| 830 |
+
{
|
| 831 |
+
"epoch": 0.34312265443497947,
|
| 832 |
+
"grad_norm": 0.08056640625,
|
| 833 |
+
"learning_rate": 0.0001661173505147295,
|
| 834 |
+
"loss": 1.091,
|
| 835 |
+
"step": 585
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"epoch": 0.34605532669510747,
|
| 839 |
+
"grad_norm": 0.07470703125,
|
| 840 |
+
"learning_rate": 0.00016534516297961996,
|
| 841 |
+
"loss": 1.0937,
|
| 842 |
+
"step": 590
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 0.3489879989552355,
|
| 846 |
+
"grad_norm": 0.0869140625,
|
| 847 |
+
"learning_rate": 0.00016456611479725996,
|
| 848 |
+
"loss": 1.1003,
|
| 849 |
+
"step": 595
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 0.3519206712153635,
|
| 853 |
+
"grad_norm": 0.07763671875,
|
| 854 |
+
"learning_rate": 0.00016378028776061667,
|
| 855 |
+
"loss": 1.0941,
|
| 856 |
+
"step": 600
|
| 857 |
+
},
|
| 858 |
+
{
|
| 859 |
+
"epoch": 0.3548533434754916,
|
| 860 |
+
"grad_norm": 0.07763671875,
|
| 861 |
+
"learning_rate": 0.00016298776437437523,
|
| 862 |
+
"loss": 1.1158,
|
| 863 |
+
"step": 605
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"epoch": 0.3577860157356196,
|
| 867 |
+
"grad_norm": 0.09228515625,
|
| 868 |
+
"learning_rate": 0.00016218862784627658,
|
| 869 |
+
"loss": 1.1004,
|
| 870 |
+
"step": 610
|
| 871 |
+
},
|
| 872 |
+
{
|
| 873 |
+
"epoch": 0.36071868799574763,
|
| 874 |
+
"grad_norm": 0.08251953125,
|
| 875 |
+
"learning_rate": 0.00016138296207838127,
|
| 876 |
+
"loss": 1.079,
|
| 877 |
+
"step": 615
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"epoch": 0.36365136025587563,
|
| 881 |
+
"grad_norm": 0.08349609375,
|
| 882 |
+
"learning_rate": 0.00016057085165826072,
|
| 883 |
+
"loss": 1.0907,
|
| 884 |
+
"step": 620
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 0.3665840325160037,
|
| 888 |
+
"grad_norm": 0.07861328125,
|
| 889 |
+
"learning_rate": 0.00015975238185011602,
|
| 890 |
+
"loss": 1.0869,
|
| 891 |
+
"step": 625
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 0.3695167047761317,
|
| 895 |
+
"grad_norm": 0.0771484375,
|
| 896 |
+
"learning_rate": 0.0001589276385858262,
|
| 897 |
+
"loss": 1.0919,
|
| 898 |
+
"step": 630
|
| 899 |
+
},
|
| 900 |
+
{
|
| 901 |
+
"epoch": 0.37244937703625974,
|
| 902 |
+
"grad_norm": 0.07421875,
|
| 903 |
+
"learning_rate": 0.00015809670845592604,
|
| 904 |
+
"loss": 1.095,
|
| 905 |
+
"step": 635
|
| 906 |
+
},
|
| 907 |
+
{
|
| 908 |
+
"epoch": 0.3753820492963878,
|
| 909 |
+
"grad_norm": 0.07470703125,
|
| 910 |
+
"learning_rate": 0.0001572596787005149,
|
| 911 |
+
"loss": 1.0989,
|
| 912 |
+
"step": 640
|
| 913 |
+
},
|
| 914 |
+
{
|
| 915 |
+
"epoch": 0.3783147215565158,
|
| 916 |
+
"grad_norm": 0.0791015625,
|
| 917 |
+
"learning_rate": 0.00015641663720009733,
|
| 918 |
+
"loss": 1.0984,
|
| 919 |
+
"step": 645
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"epoch": 0.38124739381664385,
|
| 923 |
+
"grad_norm": 0.07470703125,
|
| 924 |
+
"learning_rate": 0.00015556767246635626,
|
| 925 |
+
"loss": 1.1086,
|
| 926 |
+
"step": 650
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 0.38418006607677185,
|
| 930 |
+
"grad_norm": 0.08203125,
|
| 931 |
+
"learning_rate": 0.00015471287363286038,
|
| 932 |
+
"loss": 1.0858,
|
| 933 |
+
"step": 655
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 0.3871127383368999,
|
| 937 |
+
"grad_norm": 0.16796875,
|
| 938 |
+
"learning_rate": 0.00015385233044570555,
|
| 939 |
+
"loss": 1.096,
|
| 940 |
+
"step": 660
|
| 941 |
+
},
|
| 942 |
+
{
|
| 943 |
+
"epoch": 0.3900454105970279,
|
| 944 |
+
"grad_norm": 0.07958984375,
|
| 945 |
+
"learning_rate": 0.00015298613325409263,
|
| 946 |
+
"loss": 1.089,
|
| 947 |
+
"step": 665
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 0.39297808285715596,
|
| 951 |
+
"grad_norm": 0.07861328125,
|
| 952 |
+
"learning_rate": 0.00015211437300084136,
|
| 953 |
+
"loss": 1.0853,
|
| 954 |
+
"step": 670
|
| 955 |
+
},
|
| 956 |
+
{
|
| 957 |
+
"epoch": 0.39591075511728396,
|
| 958 |
+
"grad_norm": 0.07470703125,
|
| 959 |
+
"learning_rate": 0.0001512371412128424,
|
| 960 |
+
"loss": 1.1072,
|
| 961 |
+
"step": 675
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"epoch": 0.398843427377412,
|
| 965 |
+
"grad_norm": 0.07666015625,
|
| 966 |
+
"learning_rate": 0.00015035452999144762,
|
| 967 |
+
"loss": 1.0836,
|
| 968 |
+
"step": 680
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 0.40177609963754,
|
| 972 |
+
"grad_norm": 0.0859375,
|
| 973 |
+
"learning_rate": 0.00014946663200280063,
|
| 974 |
+
"loss": 1.0942,
|
| 975 |
+
"step": 685
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 0.4047087718976681,
|
| 979 |
+
"grad_norm": 0.072265625,
|
| 980 |
+
"learning_rate": 0.00014857354046810732,
|
| 981 |
+
"loss": 1.093,
|
| 982 |
+
"step": 690
|
| 983 |
+
},
|
| 984 |
+
{
|
| 985 |
+
"epoch": 0.4076414441577961,
|
| 986 |
+
"grad_norm": 0.0712890625,
|
| 987 |
+
"learning_rate": 0.00014767534915384865,
|
| 988 |
+
"loss": 1.0874,
|
| 989 |
+
"step": 695
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"epoch": 0.41057411641792413,
|
| 993 |
+
"grad_norm": 0.07568359375,
|
| 994 |
+
"learning_rate": 0.00014677215236193604,
|
| 995 |
+
"loss": 1.0787,
|
| 996 |
+
"step": 700
|
| 997 |
+
},
|
| 998 |
+
{
|
| 999 |
+
"epoch": 0.41350678867805213,
|
| 1000 |
+
"grad_norm": 0.07568359375,
|
| 1001 |
+
"learning_rate": 0.00014586404491981052,
|
| 1002 |
+
"loss": 1.1116,
|
| 1003 |
+
"step": 705
|
| 1004 |
+
},
|
| 1005 |
+
{
|
| 1006 |
+
"epoch": 0.4164394609381802,
|
| 1007 |
+
"grad_norm": 0.07373046875,
|
| 1008 |
+
"learning_rate": 0.00014495112217048658,
|
| 1009 |
+
"loss": 1.1018,
|
| 1010 |
+
"step": 710
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 0.41937213319830824,
|
| 1014 |
+
"grad_norm": 0.07470703125,
|
| 1015 |
+
"learning_rate": 0.00014403347996254232,
|
| 1016 |
+
"loss": 1.0945,
|
| 1017 |
+
"step": 715
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 0.42230480545843624,
|
| 1021 |
+
"grad_norm": 0.07275390625,
|
| 1022 |
+
"learning_rate": 0.00014311121464005583,
|
| 1023 |
+
"loss": 1.0836,
|
| 1024 |
+
"step": 720
|
| 1025 |
+
},
|
| 1026 |
+
{
|
| 1027 |
+
"epoch": 0.4252374777185643,
|
| 1028 |
+
"grad_norm": 0.076171875,
|
| 1029 |
+
"learning_rate": 0.00014218442303249026,
|
| 1030 |
+
"loss": 1.0765,
|
| 1031 |
+
"step": 725
|
| 1032 |
+
},
|
| 1033 |
+
{
|
| 1034 |
+
"epoch": 0.4281701499786923,
|
| 1035 |
+
"grad_norm": 0.0791015625,
|
| 1036 |
+
"learning_rate": 0.0001412532024445275,
|
| 1037 |
+
"loss": 1.0928,
|
| 1038 |
+
"step": 730
|
| 1039 |
+
},
|
| 1040 |
+
{
|
| 1041 |
+
"epoch": 0.43110282223882035,
|
| 1042 |
+
"grad_norm": 0.07763671875,
|
| 1043 |
+
"learning_rate": 0.00014031765064585197,
|
| 1044 |
+
"loss": 1.0855,
|
| 1045 |
+
"step": 735
|
| 1046 |
+
},
|
| 1047 |
+
{
|
| 1048 |
+
"epoch": 0.43403549449894835,
|
| 1049 |
+
"grad_norm": 0.07275390625,
|
| 1050 |
+
"learning_rate": 0.00013937786586088583,
|
| 1051 |
+
"loss": 1.0841,
|
| 1052 |
+
"step": 740
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 0.4369681667590764,
|
| 1056 |
+
"grad_norm": 0.07421875,
|
| 1057 |
+
"learning_rate": 0.00013843394675847634,
|
| 1058 |
+
"loss": 1.0571,
|
| 1059 |
+
"step": 745
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 0.4399008390192044,
|
| 1063 |
+
"grad_norm": 0.0810546875,
|
| 1064 |
+
"learning_rate": 0.00013748599244153633,
|
| 1065 |
+
"loss": 1.0945,
|
| 1066 |
+
"step": 750
|
| 1067 |
+
},
|
| 1068 |
+
{
|
| 1069 |
+
"epoch": 0.44283351127933246,
|
| 1070 |
+
"grad_norm": 0.0771484375,
|
| 1071 |
+
"learning_rate": 0.00013653410243663952,
|
| 1072 |
+
"loss": 1.0922,
|
| 1073 |
+
"step": 755
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 0.44576618353946046,
|
| 1077 |
+
"grad_norm": 0.0771484375,
|
| 1078 |
+
"learning_rate": 0.000135578376683571,
|
| 1079 |
+
"loss": 1.097,
|
| 1080 |
+
"step": 760
|
| 1081 |
+
},
|
| 1082 |
+
{
|
| 1083 |
+
"epoch": 0.4486988557995885,
|
| 1084 |
+
"grad_norm": 0.08056640625,
|
| 1085 |
+
"learning_rate": 0.00013461891552483444,
|
| 1086 |
+
"loss": 1.09,
|
| 1087 |
+
"step": 765
|
| 1088 |
+
},
|
| 1089 |
+
{
|
| 1090 |
+
"epoch": 0.4516315280597165,
|
| 1091 |
+
"grad_norm": 0.07666015625,
|
| 1092 |
+
"learning_rate": 0.00013365581969511725,
|
| 1093 |
+
"loss": 1.0844,
|
| 1094 |
+
"step": 770
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 0.45456420031984457,
|
| 1098 |
+
"grad_norm": 0.080078125,
|
| 1099 |
+
"learning_rate": 0.00013268919031071406,
|
| 1100 |
+
"loss": 1.099,
|
| 1101 |
+
"step": 775
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 0.4574968725799726,
|
| 1105 |
+
"grad_norm": 0.07861328125,
|
| 1106 |
+
"learning_rate": 0.00013171912885891063,
|
| 1107 |
+
"loss": 1.0919,
|
| 1108 |
+
"step": 780
|
| 1109 |
+
},
|
| 1110 |
+
{
|
| 1111 |
+
"epoch": 0.4604295448401006,
|
| 1112 |
+
"grad_norm": 0.07470703125,
|
| 1113 |
+
"learning_rate": 0.00013074573718732858,
|
| 1114 |
+
"loss": 1.0743,
|
| 1115 |
+
"step": 785
|
| 1116 |
+
},
|
| 1117 |
+
{
|
| 1118 |
+
"epoch": 0.4633622171002287,
|
| 1119 |
+
"grad_norm": 0.076171875,
|
| 1120 |
+
"learning_rate": 0.0001297691174932322,
|
| 1121 |
+
"loss": 1.0849,
|
| 1122 |
+
"step": 790
|
| 1123 |
+
},
|
| 1124 |
+
{
|
| 1125 |
+
"epoch": 0.4662948893603567,
|
| 1126 |
+
"grad_norm": 0.07421875,
|
| 1127 |
+
"learning_rate": 0.00012878937231279892,
|
| 1128 |
+
"loss": 1.0769,
|
| 1129 |
+
"step": 795
|
| 1130 |
+
},
|
| 1131 |
+
{
|
| 1132 |
+
"epoch": 0.46922756162048473,
|
| 1133 |
+
"grad_norm": 0.08154296875,
|
| 1134 |
+
"learning_rate": 0.0001278066045103536,
|
| 1135 |
+
"loss": 1.0656,
|
| 1136 |
+
"step": 800
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 0.47216023388061273,
|
| 1140 |
+
"grad_norm": 0.0791015625,
|
| 1141 |
+
"learning_rate": 0.00012682091726756904,
|
| 1142 |
+
"loss": 1.0847,
|
| 1143 |
+
"step": 805
|
| 1144 |
+
},
|
| 1145 |
+
{
|
| 1146 |
+
"epoch": 0.4750929061407408,
|
| 1147 |
+
"grad_norm": 0.078125,
|
| 1148 |
+
"learning_rate": 0.0001258324140726326,
|
| 1149 |
+
"loss": 1.0789,
|
| 1150 |
+
"step": 810
|
| 1151 |
+
},
|
| 1152 |
+
{
|
| 1153 |
+
"epoch": 0.4780255784008688,
|
| 1154 |
+
"grad_norm": 0.0732421875,
|
| 1155 |
+
"learning_rate": 0.00012484119870938103,
|
| 1156 |
+
"loss": 1.1031,
|
| 1157 |
+
"step": 815
|
| 1158 |
+
},
|
| 1159 |
+
{
|
| 1160 |
+
"epoch": 0.48095825066099684,
|
| 1161 |
+
"grad_norm": 0.0751953125,
|
| 1162 |
+
"learning_rate": 0.00012384737524640405,
|
| 1163 |
+
"loss": 1.0838,
|
| 1164 |
+
"step": 820
|
| 1165 |
+
},
|
| 1166 |
+
{
|
| 1167 |
+
"epoch": 0.48389092292112484,
|
| 1168 |
+
"grad_norm": 0.07275390625,
|
| 1169 |
+
"learning_rate": 0.00012285104802611812,
|
| 1170 |
+
"loss": 1.0795,
|
| 1171 |
+
"step": 825
|
| 1172 |
+
},
|
| 1173 |
+
{
|
| 1174 |
+
"epoch": 0.4868235951812529,
|
| 1175 |
+
"grad_norm": 0.07763671875,
|
| 1176 |
+
"learning_rate": 0.00012185232165381141,
|
| 1177 |
+
"loss": 1.0936,
|
| 1178 |
+
"step": 830
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 0.4897562674413809,
|
| 1182 |
+
"grad_norm": 0.07666015625,
|
| 1183 |
+
"learning_rate": 0.00012085130098666124,
|
| 1184 |
+
"loss": 1.0845,
|
| 1185 |
+
"step": 835
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 0.49268893970150895,
|
| 1189 |
+
"grad_norm": 0.09130859375,
|
| 1190 |
+
"learning_rate": 0.00011984809112272495,
|
| 1191 |
+
"loss": 1.0807,
|
| 1192 |
+
"step": 840
|
| 1193 |
+
},
|
| 1194 |
+
{
|
| 1195 |
+
"epoch": 0.49562161196163695,
|
| 1196 |
+
"grad_norm": 0.07666015625,
|
| 1197 |
+
"learning_rate": 0.00011884279738990565,
|
| 1198 |
+
"loss": 1.0834,
|
| 1199 |
+
"step": 845
|
| 1200 |
+
},
|
| 1201 |
+
{
|
| 1202 |
+
"epoch": 0.498554284221765,
|
| 1203 |
+
"grad_norm": 0.08203125,
|
| 1204 |
+
"learning_rate": 0.00011783552533489372,
|
| 1205 |
+
"loss": 1.0944,
|
| 1206 |
+
"step": 850
|
| 1207 |
+
},
|
| 1208 |
+
{
|
| 1209 |
+
"epoch": 0.501486956481893,
|
| 1210 |
+
"grad_norm": 0.0771484375,
|
| 1211 |
+
"learning_rate": 0.00011682638071208533,
|
| 1212 |
+
"loss": 1.095,
|
| 1213 |
+
"step": 855
|
| 1214 |
+
},
|
| 1215 |
+
{
|
| 1216 |
+
"epoch": 0.5044196287420211,
|
| 1217 |
+
"grad_norm": 0.08056640625,
|
| 1218 |
+
"learning_rate": 0.00011581546947247927,
|
| 1219 |
+
"loss": 1.0819,
|
| 1220 |
+
"step": 860
|
| 1221 |
+
},
|
| 1222 |
+
{
|
| 1223 |
+
"epoch": 0.5073523010021491,
|
| 1224 |
+
"grad_norm": 0.0732421875,
|
| 1225 |
+
"learning_rate": 0.00011480289775255295,
|
| 1226 |
+
"loss": 1.0875,
|
| 1227 |
+
"step": 865
|
| 1228 |
+
},
|
| 1229 |
+
{
|
| 1230 |
+
"epoch": 0.5102849732622772,
|
| 1231 |
+
"grad_norm": 0.07666015625,
|
| 1232 |
+
"learning_rate": 0.00011378877186311912,
|
| 1233 |
+
"loss": 1.0879,
|
| 1234 |
+
"step": 870
|
| 1235 |
+
},
|
| 1236 |
+
{
|
| 1237 |
+
"epoch": 0.5132176455224051,
|
| 1238 |
+
"grad_norm": 0.0751953125,
|
| 1239 |
+
"learning_rate": 0.00011277319827816423,
|
| 1240 |
+
"loss": 1.1074,
|
| 1241 |
+
"step": 875
|
| 1242 |
+
},
|
| 1243 |
+
{
|
| 1244 |
+
"epoch": 0.5161503177825332,
|
| 1245 |
+
"grad_norm": 0.0771484375,
|
| 1246 |
+
"learning_rate": 0.0001117562836236695,
|
| 1247 |
+
"loss": 1.0842,
|
| 1248 |
+
"step": 880
|
| 1249 |
+
},
|
| 1250 |
+
{
|
| 1251 |
+
"epoch": 0.5190829900426612,
|
| 1252 |
+
"grad_norm": 0.078125,
|
| 1253 |
+
"learning_rate": 0.00011073813466641632,
|
| 1254 |
+
"loss": 1.0812,
|
| 1255 |
+
"step": 885
|
| 1256 |
+
},
|
| 1257 |
+
{
|
| 1258 |
+
"epoch": 0.5220156623027893,
|
| 1259 |
+
"grad_norm": 0.0732421875,
|
| 1260 |
+
"learning_rate": 0.00010971885830277657,
|
| 1261 |
+
"loss": 1.108,
|
| 1262 |
+
"step": 890
|
| 1263 |
+
},
|
| 1264 |
+
{
|
| 1265 |
+
"epoch": 0.5249483345629172,
|
| 1266 |
+
"grad_norm": 0.0849609375,
|
| 1267 |
+
"learning_rate": 0.00010869856154748956,
|
| 1268 |
+
"loss": 1.0808,
|
| 1269 |
+
"step": 895
|
| 1270 |
+
},
|
| 1271 |
+
{
|
| 1272 |
+
"epoch": 0.5278810068230453,
|
| 1273 |
+
"grad_norm": 0.076171875,
|
| 1274 |
+
"learning_rate": 0.00010767735152242649,
|
| 1275 |
+
"loss": 1.0734,
|
| 1276 |
+
"step": 900
|
| 1277 |
+
},
|
| 1278 |
+
{
|
| 1279 |
+
"epoch": 0.5308136790831733,
|
| 1280 |
+
"grad_norm": 0.07470703125,
|
| 1281 |
+
"learning_rate": 0.00010665533544534343,
|
| 1282 |
+
"loss": 1.0865,
|
| 1283 |
+
"step": 905
|
| 1284 |
+
},
|
| 1285 |
+
{
|
| 1286 |
+
"epoch": 0.5337463513433014,
|
| 1287 |
+
"grad_norm": 0.078125,
|
| 1288 |
+
"learning_rate": 0.00010563262061862471,
|
| 1289 |
+
"loss": 1.0818,
|
| 1290 |
+
"step": 910
|
| 1291 |
+
},
|
| 1292 |
+
{
|
| 1293 |
+
"epoch": 0.5366790236034293,
|
| 1294 |
+
"grad_norm": 0.07421875,
|
| 1295 |
+
"learning_rate": 0.000104609314418017,
|
| 1296 |
+
"loss": 1.0719,
|
| 1297 |
+
"step": 915
|
| 1298 |
+
},
|
| 1299 |
+
{
|
| 1300 |
+
"epoch": 0.5396116958635574,
|
| 1301 |
+
"grad_norm": 0.07763671875,
|
| 1302 |
+
"learning_rate": 0.00010358552428135575,
|
| 1303 |
+
"loss": 1.0836,
|
| 1304 |
+
"step": 920
|
| 1305 |
+
},
|
| 1306 |
+
{
|
| 1307 |
+
"epoch": 0.5425443681236854,
|
| 1308 |
+
"grad_norm": 0.07421875,
|
| 1309 |
+
"learning_rate": 0.00010256135769728539,
|
| 1310 |
+
"loss": 1.0967,
|
| 1311 |
+
"step": 925
|
| 1312 |
+
},
|
| 1313 |
+
{
|
| 1314 |
+
"epoch": 0.5454770403838135,
|
| 1315 |
+
"grad_norm": 0.0732421875,
|
| 1316 |
+
"learning_rate": 0.00010153692219397387,
|
| 1317 |
+
"loss": 1.0691,
|
| 1318 |
+
"step": 930
|
| 1319 |
+
},
|
| 1320 |
+
{
|
| 1321 |
+
"epoch": 0.5484097126439416,
|
| 1322 |
+
"grad_norm": 0.07275390625,
|
| 1323 |
+
"learning_rate": 0.00010051232532782313,
|
| 1324 |
+
"loss": 1.0766,
|
| 1325 |
+
"step": 935
|
| 1326 |
+
},
|
| 1327 |
+
{
|
| 1328 |
+
"epoch": 0.5513423849040695,
|
| 1329 |
+
"grad_norm": 0.07666015625,
|
| 1330 |
+
"learning_rate": 9.94876746721769e-05,
|
| 1331 |
+
"loss": 1.0727,
|
| 1332 |
+
"step": 940
|
| 1333 |
+
},
|
| 1334 |
+
{
|
| 1335 |
+
"epoch": 0.5542750571641976,
|
| 1336 |
+
"grad_norm": 0.07568359375,
|
| 1337 |
+
"learning_rate": 9.84630778060262e-05,
|
| 1338 |
+
"loss": 1.0784,
|
| 1339 |
+
"step": 945
|
| 1340 |
+
},
|
| 1341 |
+
{
|
| 1342 |
+
"epoch": 0.5572077294243256,
|
| 1343 |
+
"grad_norm": 0.0791015625,
|
| 1344 |
+
"learning_rate": 9.743864230271465e-05,
|
| 1345 |
+
"loss": 1.0809,
|
| 1346 |
+
"step": 950
|
| 1347 |
+
},
|
| 1348 |
+
{
|
| 1349 |
+
"epoch": 0.5601404016844537,
|
| 1350 |
+
"grad_norm": 0.078125,
|
| 1351 |
+
"learning_rate": 9.641447571864429e-05,
|
| 1352 |
+
"loss": 1.0809,
|
| 1353 |
+
"step": 955
|
| 1354 |
+
},
|
| 1355 |
+
{
|
| 1356 |
+
"epoch": 0.5630730739445816,
|
| 1357 |
+
"grad_norm": 0.07470703125,
|
| 1358 |
+
"learning_rate": 9.539068558198304e-05,
|
| 1359 |
+
"loss": 1.0784,
|
| 1360 |
+
"step": 960
|
| 1361 |
+
},
|
| 1362 |
+
{
|
| 1363 |
+
"epoch": 0.5660057462047097,
|
| 1364 |
+
"grad_norm": 0.0751953125,
|
| 1365 |
+
"learning_rate": 9.436737938137531e-05,
|
| 1366 |
+
"loss": 1.0707,
|
| 1367 |
+
"step": 965
|
| 1368 |
+
},
|
| 1369 |
+
{
|
| 1370 |
+
"epoch": 0.5689384184648377,
|
| 1371 |
+
"grad_norm": 0.0751953125,
|
| 1372 |
+
"learning_rate": 9.33446645546566e-05,
|
| 1373 |
+
"loss": 1.0821,
|
| 1374 |
+
"step": 970
|
| 1375 |
+
},
|
| 1376 |
+
{
|
| 1377 |
+
"epoch": 0.5718710907249658,
|
| 1378 |
+
"grad_norm": 0.07666015625,
|
| 1379 |
+
"learning_rate": 9.232264847757357e-05,
|
| 1380 |
+
"loss": 1.0715,
|
| 1381 |
+
"step": 975
|
| 1382 |
+
},
|
| 1383 |
+
{
|
| 1384 |
+
"epoch": 0.5748037629850937,
|
| 1385 |
+
"grad_norm": 0.078125,
|
| 1386 |
+
"learning_rate": 9.130143845251046e-05,
|
| 1387 |
+
"loss": 1.0746,
|
| 1388 |
+
"step": 980
|
| 1389 |
+
},
|
| 1390 |
+
{
|
| 1391 |
+
"epoch": 0.5777364352452218,
|
| 1392 |
+
"grad_norm": 0.0732421875,
|
| 1393 |
+
"learning_rate": 9.028114169722347e-05,
|
| 1394 |
+
"loss": 1.0987,
|
| 1395 |
+
"step": 985
|
| 1396 |
+
},
|
| 1397 |
+
{
|
| 1398 |
+
"epoch": 0.5806691075053498,
|
| 1399 |
+
"grad_norm": 0.07666015625,
|
| 1400 |
+
"learning_rate": 8.92618653335837e-05,
|
| 1401 |
+
"loss": 1.0758,
|
| 1402 |
+
"step": 990
|
| 1403 |
+
},
|
| 1404 |
+
{
|
| 1405 |
+
"epoch": 0.5836017797654779,
|
| 1406 |
+
"grad_norm": 0.07373046875,
|
| 1407 |
+
"learning_rate": 8.824371637633053e-05,
|
| 1408 |
+
"loss": 1.0523,
|
| 1409 |
+
"step": 995
|
| 1410 |
+
},
|
| 1411 |
+
{
|
| 1412 |
+
"epoch": 0.586534452025606,
|
| 1413 |
+
"grad_norm": 0.07666015625,
|
| 1414 |
+
"learning_rate": 8.722680172183578e-05,
|
| 1415 |
+
"loss": 1.0797,
|
| 1416 |
+
"step": 1000
|
| 1417 |
+
},
|
| 1418 |
+
{
|
| 1419 |
+
"epoch": 0.5894671242857339,
|
| 1420 |
+
"grad_norm": 0.07470703125,
|
| 1421 |
+
"learning_rate": 8.62112281368809e-05,
|
| 1422 |
+
"loss": 1.0951,
|
| 1423 |
+
"step": 1005
|
| 1424 |
+
},
|
| 1425 |
+
{
|
| 1426 |
+
"epoch": 0.592399796545862,
|
| 1427 |
+
"grad_norm": 0.076171875,
|
| 1428 |
+
"learning_rate": 8.519710224744709e-05,
|
| 1429 |
+
"loss": 1.0833,
|
| 1430 |
+
"step": 1010
|
| 1431 |
+
},
|
| 1432 |
+
{
|
| 1433 |
+
"epoch": 0.59533246880599,
|
| 1434 |
+
"grad_norm": 0.07861328125,
|
| 1435 |
+
"learning_rate": 8.418453052752076e-05,
|
| 1436 |
+
"loss": 1.0717,
|
| 1437 |
+
"step": 1015
|
| 1438 |
+
},
|
| 1439 |
+
{
|
| 1440 |
+
"epoch": 0.5982651410661181,
|
| 1441 |
+
"grad_norm": 0.078125,
|
| 1442 |
+
"learning_rate": 8.317361928791469e-05,
|
| 1443 |
+
"loss": 1.0757,
|
| 1444 |
+
"step": 1020
|
| 1445 |
+
},
|
| 1446 |
+
{
|
| 1447 |
+
"epoch": 0.601197813326246,
|
| 1448 |
+
"grad_norm": 0.07861328125,
|
| 1449 |
+
"learning_rate": 8.216447466510631e-05,
|
| 1450 |
+
"loss": 1.0819,
|
| 1451 |
+
"step": 1025
|
| 1452 |
+
},
|
| 1453 |
+
{
|
| 1454 |
+
"epoch": 0.604130485586374,
|
| 1455 |
+
"grad_norm": 0.076171875,
|
| 1456 |
+
"learning_rate": 8.115720261009437e-05,
|
| 1457 |
+
"loss": 1.075,
|
| 1458 |
+
"step": 1030
|
| 1459 |
+
},
|
| 1460 |
+
{
|
| 1461 |
+
"epoch": 0.6070631578465021,
|
| 1462 |
+
"grad_norm": 0.0751953125,
|
| 1463 |
+
"learning_rate": 8.015190887727509e-05,
|
| 1464 |
+
"loss": 1.0778,
|
| 1465 |
+
"step": 1035
|
| 1466 |
+
},
|
| 1467 |
+
{
|
| 1468 |
+
"epoch": 0.6099958301066302,
|
| 1469 |
+
"grad_norm": 0.07666015625,
|
| 1470 |
+
"learning_rate": 7.914869901333877e-05,
|
| 1471 |
+
"loss": 1.0564,
|
| 1472 |
+
"step": 1040
|
| 1473 |
+
},
|
| 1474 |
+
{
|
| 1475 |
+
"epoch": 0.6129285023667581,
|
| 1476 |
+
"grad_norm": 0.07470703125,
|
| 1477 |
+
"learning_rate": 7.81476783461886e-05,
|
| 1478 |
+
"loss": 1.0812,
|
| 1479 |
+
"step": 1045
|
| 1480 |
+
},
|
| 1481 |
+
{
|
| 1482 |
+
"epoch": 0.6158611746268862,
|
| 1483 |
+
"grad_norm": 0.0751953125,
|
| 1484 |
+
"learning_rate": 7.714895197388189e-05,
|
| 1485 |
+
"loss": 1.0919,
|
| 1486 |
+
"step": 1050
|
| 1487 |
+
},
|
| 1488 |
+
{
|
| 1489 |
+
"epoch": 0.6187938468870142,
|
| 1490 |
+
"grad_norm": 0.0732421875,
|
| 1491 |
+
"learning_rate": 7.615262475359597e-05,
|
| 1492 |
+
"loss": 1.0833,
|
| 1493 |
+
"step": 1055
|
| 1494 |
+
},
|
| 1495 |
+
{
|
| 1496 |
+
"epoch": 0.6217265191471423,
|
| 1497 |
+
"grad_norm": 0.07275390625,
|
| 1498 |
+
"learning_rate": 7.5158801290619e-05,
|
| 1499 |
+
"loss": 1.081,
|
| 1500 |
+
"step": 1060
|
| 1501 |
+
},
|
| 1502 |
+
{
|
| 1503 |
+
"epoch": 0.6246591914072703,
|
| 1504 |
+
"grad_norm": 0.07568359375,
|
| 1505 |
+
"learning_rate": 7.416758592736744e-05,
|
| 1506 |
+
"loss": 1.0686,
|
| 1507 |
+
"step": 1065
|
| 1508 |
+
},
|
| 1509 |
+
{
|
| 1510 |
+
"epoch": 0.6275918636673983,
|
| 1511 |
+
"grad_norm": 0.0732421875,
|
| 1512 |
+
"learning_rate": 7.3179082732431e-05,
|
| 1513 |
+
"loss": 1.0728,
|
| 1514 |
+
"step": 1070
|
| 1515 |
+
},
|
| 1516 |
+
{
|
| 1517 |
+
"epoch": 0.6305245359275263,
|
| 1518 |
+
"grad_norm": 0.0771484375,
|
| 1519 |
+
"learning_rate": 7.219339548964644e-05,
|
| 1520 |
+
"loss": 1.0735,
|
| 1521 |
+
"step": 1075
|
| 1522 |
+
},
|
| 1523 |
+
{
|
| 1524 |
+
"epoch": 0.6334572081876544,
|
| 1525 |
+
"grad_norm": 0.0771484375,
|
| 1526 |
+
"learning_rate": 7.12106276872011e-05,
|
| 1527 |
+
"loss": 1.0621,
|
| 1528 |
+
"step": 1080
|
| 1529 |
+
},
|
| 1530 |
+
{
|
| 1531 |
+
"epoch": 0.6363898804477824,
|
| 1532 |
+
"grad_norm": 0.0751953125,
|
| 1533 |
+
"learning_rate": 7.023088250676784e-05,
|
| 1534 |
+
"loss": 1.071,
|
| 1535 |
+
"step": 1085
|
| 1536 |
+
},
|
| 1537 |
+
{
|
| 1538 |
+
"epoch": 0.6393225527079104,
|
| 1539 |
+
"grad_norm": 0.07763671875,
|
| 1540 |
+
"learning_rate": 6.925426281267147e-05,
|
| 1541 |
+
"loss": 1.0794,
|
| 1542 |
+
"step": 1090
|
| 1543 |
+
},
|
| 1544 |
+
{
|
| 1545 |
+
"epoch": 0.6422552249680384,
|
| 1546 |
+
"grad_norm": 0.0791015625,
|
| 1547 |
+
"learning_rate": 6.82808711410894e-05,
|
| 1548 |
+
"loss": 1.0637,
|
| 1549 |
+
"step": 1095
|
| 1550 |
+
},
|
| 1551 |
+
{
|
| 1552 |
+
"epoch": 0.6451878972281665,
|
| 1553 |
+
"grad_norm": 0.07275390625,
|
| 1554 |
+
"learning_rate": 6.731080968928599e-05,
|
| 1555 |
+
"loss": 1.0665,
|
| 1556 |
+
"step": 1100
|
| 1557 |
+
},
|
| 1558 |
+
{
|
| 1559 |
+
"epoch": 0.6481205694882946,
|
| 1560 |
+
"grad_norm": 0.0732421875,
|
| 1561 |
+
"learning_rate": 6.63441803048828e-05,
|
| 1562 |
+
"loss": 1.0773,
|
| 1563 |
+
"step": 1105
|
| 1564 |
+
},
|
| 1565 |
+
{
|
| 1566 |
+
"epoch": 0.6510532417484225,
|
| 1567 |
+
"grad_norm": 0.07470703125,
|
| 1568 |
+
"learning_rate": 6.538108447516558e-05,
|
| 1569 |
+
"loss": 1.0689,
|
| 1570 |
+
"step": 1110
|
| 1571 |
+
},
|
| 1572 |
+
{
|
| 1573 |
+
"epoch": 0.6539859140085506,
|
| 1574 |
+
"grad_norm": 0.0751953125,
|
| 1575 |
+
"learning_rate": 6.442162331642907e-05,
|
| 1576 |
+
"loss": 1.0894,
|
| 1577 |
+
"step": 1115
|
| 1578 |
+
},
|
| 1579 |
+
{
|
| 1580 |
+
"epoch": 0.6569185862686786,
|
| 1581 |
+
"grad_norm": 0.072265625,
|
| 1582 |
+
"learning_rate": 6.34658975633605e-05,
|
| 1583 |
+
"loss": 1.0772,
|
| 1584 |
+
"step": 1120
|
| 1585 |
+
},
|
| 1586 |
+
{
|
| 1587 |
+
"epoch": 0.6598512585288067,
|
| 1588 |
+
"grad_norm": 0.07470703125,
|
| 1589 |
+
"learning_rate": 6.251400755846372e-05,
|
| 1590 |
+
"loss": 1.0653,
|
| 1591 |
+
"step": 1125
|
| 1592 |
+
},
|
| 1593 |
+
{
|
| 1594 |
+
"epoch": 0.6627839307889347,
|
| 1595 |
+
"grad_norm": 0.07568359375,
|
| 1596 |
+
"learning_rate": 6.15660532415237e-05,
|
| 1597 |
+
"loss": 1.0827,
|
| 1598 |
+
"step": 1130
|
| 1599 |
+
},
|
| 1600 |
+
{
|
| 1601 |
+
"epoch": 0.6657166030490627,
|
| 1602 |
+
"grad_norm": 0.076171875,
|
| 1603 |
+
"learning_rate": 6.0622134139114194e-05,
|
| 1604 |
+
"loss": 1.0833,
|
| 1605 |
+
"step": 1135
|
| 1606 |
+
},
|
| 1607 |
+
{
|
| 1608 |
+
"epoch": 0.6686492753091907,
|
| 1609 |
+
"grad_norm": 0.07373046875,
|
| 1610 |
+
"learning_rate": 5.968234935414807e-05,
|
| 1611 |
+
"loss": 1.081,
|
| 1612 |
+
"step": 1140
|
| 1613 |
+
},
|
| 1614 |
+
{
|
| 1615 |
+
"epoch": 0.6715819475693188,
|
| 1616 |
+
"grad_norm": 0.0771484375,
|
| 1617 |
+
"learning_rate": 5.874679755547254e-05,
|
| 1618 |
+
"loss": 1.0955,
|
| 1619 |
+
"step": 1145
|
| 1620 |
+
},
|
| 1621 |
+
{
|
| 1622 |
+
"epoch": 0.6745146198294468,
|
| 1623 |
+
"grad_norm": 0.07470703125,
|
| 1624 |
+
"learning_rate": 5.7815576967509733e-05,
|
| 1625 |
+
"loss": 1.0862,
|
| 1626 |
+
"step": 1150
|
| 1627 |
+
},
|
| 1628 |
+
{
|
| 1629 |
+
"epoch": 0.6774472920895748,
|
| 1630 |
+
"grad_norm": 0.07763671875,
|
| 1631 |
+
"learning_rate": 5.688878535994421e-05,
|
| 1632 |
+
"loss": 1.0885,
|
| 1633 |
+
"step": 1155
|
| 1634 |
+
},
|
| 1635 |
+
{
|
| 1636 |
+
"epoch": 0.6803799643497028,
|
| 1637 |
+
"grad_norm": 0.078125,
|
| 1638 |
+
"learning_rate": 5.5966520037457716e-05,
|
| 1639 |
+
"loss": 1.0578,
|
| 1640 |
+
"step": 1160
|
| 1641 |
+
},
|
| 1642 |
+
{
|
| 1643 |
+
"epoch": 0.6833126366098309,
|
| 1644 |
+
"grad_norm": 0.0751953125,
|
| 1645 |
+
"learning_rate": 5.5048877829513424e-05,
|
| 1646 |
+
"loss": 1.0914,
|
| 1647 |
+
"step": 1165
|
| 1648 |
+
},
|
| 1649 |
+
{
|
| 1650 |
+
"epoch": 0.6862453088699589,
|
| 1651 |
+
"grad_norm": 0.07470703125,
|
| 1652 |
+
"learning_rate": 5.413595508018952e-05,
|
| 1653 |
+
"loss": 1.0837,
|
| 1654 |
+
"step": 1170
|
| 1655 |
+
},
|
| 1656 |
+
{
|
| 1657 |
+
"epoch": 0.6891779811300869,
|
| 1658 |
+
"grad_norm": 0.07568359375,
|
| 1659 |
+
"learning_rate": 5.3227847638064e-05,
|
| 1660 |
+
"loss": 1.0665,
|
| 1661 |
+
"step": 1175
|
| 1662 |
+
},
|
| 1663 |
+
{
|
| 1664 |
+
"epoch": 0.6921106533902149,
|
| 1665 |
+
"grad_norm": 0.078125,
|
| 1666 |
+
"learning_rate": 5.232465084615135e-05,
|
| 1667 |
+
"loss": 1.0815,
|
| 1668 |
+
"step": 1180
|
| 1669 |
+
},
|
| 1670 |
+
{
|
| 1671 |
+
"epoch": 0.695043325650343,
|
| 1672 |
+
"grad_norm": 0.07470703125,
|
| 1673 |
+
"learning_rate": 5.1426459531892714e-05,
|
| 1674 |
+
"loss": 1.1066,
|
| 1675 |
+
"step": 1185
|
| 1676 |
+
},
|
| 1677 |
+
{
|
| 1678 |
+
"epoch": 0.697975997910471,
|
| 1679 |
+
"grad_norm": 0.07666015625,
|
| 1680 |
+
"learning_rate": 5.0533367997199376e-05,
|
| 1681 |
+
"loss": 1.0758,
|
| 1682 |
+
"step": 1190
|
| 1683 |
+
},
|
| 1684 |
+
{
|
| 1685 |
+
"epoch": 0.700908670170599,
|
| 1686 |
+
"grad_norm": 0.076171875,
|
| 1687 |
+
"learning_rate": 4.964547000855237e-05,
|
| 1688 |
+
"loss": 1.0704,
|
| 1689 |
+
"step": 1195
|
| 1690 |
+
},
|
| 1691 |
+
{
|
| 1692 |
+
"epoch": 0.703841342430727,
|
| 1693 |
+
"grad_norm": 0.07470703125,
|
| 1694 |
+
"learning_rate": 4.876285878715764e-05,
|
| 1695 |
+
"loss": 1.067,
|
| 1696 |
+
"step": 1200
|
| 1697 |
+
},
|
| 1698 |
+
{
|
| 1699 |
+
"epoch": 0.7067740146908551,
|
| 1700 |
+
"grad_norm": 0.078125,
|
| 1701 |
+
"learning_rate": 4.7885626999158695e-05,
|
| 1702 |
+
"loss": 1.0796,
|
| 1703 |
+
"step": 1205
|
| 1704 |
+
},
|
| 1705 |
+
{
|
| 1706 |
+
"epoch": 0.7097066869509832,
|
| 1707 |
+
"grad_norm": 0.07568359375,
|
| 1708 |
+
"learning_rate": 4.701386674590742e-05,
|
| 1709 |
+
"loss": 1.0572,
|
| 1710 |
+
"step": 1210
|
| 1711 |
+
},
|
| 1712 |
+
{
|
| 1713 |
+
"epoch": 0.7126393592111112,
|
| 1714 |
+
"grad_norm": 0.07666015625,
|
| 1715 |
+
"learning_rate": 4.614766955429447e-05,
|
| 1716 |
+
"loss": 1.0896,
|
| 1717 |
+
"step": 1215
|
| 1718 |
+
},
|
| 1719 |
+
{
|
| 1720 |
+
"epoch": 0.7155720314712392,
|
| 1721 |
+
"grad_norm": 0.07568359375,
|
| 1722 |
+
"learning_rate": 4.528712636713964e-05,
|
| 1723 |
+
"loss": 1.0699,
|
| 1724 |
+
"step": 1220
|
| 1725 |
+
},
|
| 1726 |
+
{
|
| 1727 |
+
"epoch": 0.7185047037313672,
|
| 1728 |
+
"grad_norm": 0.07470703125,
|
| 1729 |
+
"learning_rate": 4.443232753364377e-05,
|
| 1730 |
+
"loss": 1.081,
|
| 1731 |
+
"step": 1225
|
| 1732 |
+
},
|
| 1733 |
+
{
|
| 1734 |
+
"epoch": 0.7214373759914953,
|
| 1735 |
+
"grad_norm": 0.0751953125,
|
| 1736 |
+
"learning_rate": 4.358336279990268e-05,
|
| 1737 |
+
"loss": 1.0733,
|
| 1738 |
+
"step": 1230
|
| 1739 |
+
},
|
| 1740 |
+
{
|
| 1741 |
+
"epoch": 0.7243700482516233,
|
| 1742 |
+
"grad_norm": 0.07666015625,
|
| 1743 |
+
"learning_rate": 4.274032129948512e-05,
|
| 1744 |
+
"loss": 1.0983,
|
| 1745 |
+
"step": 1235
|
| 1746 |
+
},
|
| 1747 |
+
{
|
| 1748 |
+
"epoch": 0.7273027205117513,
|
| 1749 |
+
"grad_norm": 0.07861328125,
|
| 1750 |
+
"learning_rate": 4.1903291544073986e-05,
|
| 1751 |
+
"loss": 1.0868,
|
| 1752 |
+
"step": 1240
|
| 1753 |
+
},
|
| 1754 |
+
{
|
| 1755 |
+
"epoch": 0.7302353927718793,
|
| 1756 |
+
"grad_norm": 0.07861328125,
|
| 1757 |
+
"learning_rate": 4.107236141417382e-05,
|
| 1758 |
+
"loss": 1.0871,
|
| 1759 |
+
"step": 1245
|
| 1760 |
+
},
|
| 1761 |
+
{
|
| 1762 |
+
"epoch": 0.7331680650320074,
|
| 1763 |
+
"grad_norm": 0.07373046875,
|
| 1764 |
+
"learning_rate": 4.024761814988399e-05,
|
| 1765 |
+
"loss": 1.0787,
|
| 1766 |
+
"step": 1250
|
| 1767 |
+
},
|
| 1768 |
+
{
|
| 1769 |
+
"epoch": 0.7361007372921354,
|
| 1770 |
+
"grad_norm": 0.07421875,
|
| 1771 |
+
"learning_rate": 3.942914834173932e-05,
|
| 1772 |
+
"loss": 1.0823,
|
| 1773 |
+
"step": 1255
|
| 1774 |
+
},
|
| 1775 |
+
{
|
| 1776 |
+
"epoch": 0.7390334095522634,
|
| 1777 |
+
"grad_norm": 0.0771484375,
|
| 1778 |
+
"learning_rate": 3.8617037921618705e-05,
|
| 1779 |
+
"loss": 1.0632,
|
| 1780 |
+
"step": 1260
|
| 1781 |
+
},
|
| 1782 |
+
{
|
| 1783 |
+
"epoch": 0.7419660818123914,
|
| 1784 |
+
"grad_norm": 0.07470703125,
|
| 1785 |
+
"learning_rate": 3.781137215372345e-05,
|
| 1786 |
+
"loss": 1.0764,
|
| 1787 |
+
"step": 1265
|
| 1788 |
+
},
|
| 1789 |
+
{
|
| 1790 |
+
"epoch": 0.7448987540725195,
|
| 1791 |
+
"grad_norm": 0.07421875,
|
| 1792 |
+
"learning_rate": 3.701223562562478e-05,
|
| 1793 |
+
"loss": 1.0878,
|
| 1794 |
+
"step": 1270
|
| 1795 |
+
},
|
| 1796 |
+
{
|
| 1797 |
+
"epoch": 0.7478314263326475,
|
| 1798 |
+
"grad_norm": 0.07373046875,
|
| 1799 |
+
"learning_rate": 3.621971223938334e-05,
|
| 1800 |
+
"loss": 1.0832,
|
| 1801 |
+
"step": 1275
|
| 1802 |
+
},
|
| 1803 |
+
{
|
| 1804 |
+
"epoch": 0.7507640985927756,
|
| 1805 |
+
"grad_norm": 0.076171875,
|
| 1806 |
+
"learning_rate": 3.5433885202740045e-05,
|
| 1807 |
+
"loss": 1.0745,
|
| 1808 |
+
"step": 1280
|
| 1809 |
+
},
|
| 1810 |
+
{
|
| 1811 |
+
"epoch": 0.7536967708529035,
|
| 1812 |
+
"grad_norm": 0.07470703125,
|
| 1813 |
+
"learning_rate": 3.4654837020380074e-05,
|
| 1814 |
+
"loss": 1.0719,
|
| 1815 |
+
"step": 1285
|
| 1816 |
+
},
|
| 1817 |
+
{
|
| 1818 |
+
"epoch": 0.7566294431130316,
|
| 1819 |
+
"grad_norm": 0.07373046875,
|
| 1820 |
+
"learning_rate": 3.388264948527052e-05,
|
| 1821 |
+
"loss": 1.0929,
|
| 1822 |
+
"step": 1290
|
| 1823 |
+
},
|
| 1824 |
+
{
|
| 1825 |
+
"epoch": 0.7595621153731597,
|
| 1826 |
+
"grad_norm": 0.07373046875,
|
| 1827 |
+
"learning_rate": 3.311740367007311e-05,
|
| 1828 |
+
"loss": 1.067,
|
| 1829 |
+
"step": 1295
|
| 1830 |
+
},
|
| 1831 |
+
{
|
| 1832 |
+
"epoch": 0.7624947876332877,
|
| 1833 |
+
"grad_norm": 0.07421875,
|
| 1834 |
+
"learning_rate": 3.2359179918632076e-05,
|
| 1835 |
+
"loss": 1.062,
|
| 1836 |
+
"step": 1300
|
| 1837 |
+
},
|
| 1838 |
+
{
|
| 1839 |
+
"epoch": 0.7654274598934157,
|
| 1840 |
+
"grad_norm": 0.07275390625,
|
| 1841 |
+
"learning_rate": 3.160805783753897e-05,
|
| 1842 |
+
"loss": 1.06,
|
| 1843 |
+
"step": 1305
|
| 1844 |
+
},
|
| 1845 |
+
{
|
| 1846 |
+
"epoch": 0.7683601321535437,
|
| 1847 |
+
"grad_norm": 0.0751953125,
|
| 1848 |
+
"learning_rate": 3.086411628777445e-05,
|
| 1849 |
+
"loss": 1.084,
|
| 1850 |
+
"step": 1310
|
| 1851 |
+
},
|
| 1852 |
+
{
|
| 1853 |
+
"epoch": 0.7712928044136718,
|
| 1854 |
+
"grad_norm": 0.078125,
|
| 1855 |
+
"learning_rate": 3.0127433376428983e-05,
|
| 1856 |
+
"loss": 1.0578,
|
| 1857 |
+
"step": 1315
|
| 1858 |
+
},
|
| 1859 |
+
{
|
| 1860 |
+
"epoch": 0.7742254766737998,
|
| 1861 |
+
"grad_norm": 0.0732421875,
|
| 1862 |
+
"learning_rate": 2.939808644850184e-05,
|
| 1863 |
+
"loss": 1.0637,
|
| 1864 |
+
"step": 1320
|
| 1865 |
+
},
|
| 1866 |
+
{
|
| 1867 |
+
"epoch": 0.7771581489339278,
|
| 1868 |
+
"grad_norm": 0.07177734375,
|
| 1869 |
+
"learning_rate": 2.867615207878096e-05,
|
| 1870 |
+
"loss": 1.0661,
|
| 1871 |
+
"step": 1325
|
| 1872 |
+
},
|
| 1873 |
+
{
|
| 1874 |
+
"epoch": 0.7800908211940558,
|
| 1875 |
+
"grad_norm": 0.0751953125,
|
| 1876 |
+
"learning_rate": 2.796170606380313e-05,
|
| 1877 |
+
"loss": 1.0768,
|
| 1878 |
+
"step": 1330
|
| 1879 |
+
},
|
| 1880 |
+
{
|
| 1881 |
+
"epoch": 0.7830234934541839,
|
| 1882 |
+
"grad_norm": 0.07470703125,
|
| 1883 |
+
"learning_rate": 2.7254823413896058e-05,
|
| 1884 |
+
"loss": 1.068,
|
| 1885 |
+
"step": 1335
|
| 1886 |
+
},
|
| 1887 |
+
{
|
| 1888 |
+
"epoch": 0.7859561657143119,
|
| 1889 |
+
"grad_norm": 0.07470703125,
|
| 1890 |
+
"learning_rate": 2.6555578345302878e-05,
|
| 1891 |
+
"loss": 1.0674,
|
| 1892 |
+
"step": 1340
|
| 1893 |
+
},
|
| 1894 |
+
{
|
| 1895 |
+
"epoch": 0.78888883797444,
|
| 1896 |
+
"grad_norm": 0.07373046875,
|
| 1897 |
+
"learning_rate": 2.58640442723904e-05,
|
| 1898 |
+
"loss": 1.0798,
|
| 1899 |
+
"step": 1345
|
| 1900 |
+
},
|
| 1901 |
+
{
|
| 1902 |
+
"epoch": 0.7918215102345679,
|
| 1903 |
+
"grad_norm": 0.0751953125,
|
| 1904 |
+
"learning_rate": 2.518029379994089e-05,
|
| 1905 |
+
"loss": 1.0694,
|
| 1906 |
+
"step": 1350
|
| 1907 |
+
},
|
| 1908 |
+
{
|
| 1909 |
+
"epoch": 0.794754182494696,
|
| 1910 |
+
"grad_norm": 0.0771484375,
|
| 1911 |
+
"learning_rate": 2.4504398715529554e-05,
|
| 1912 |
+
"loss": 1.0752,
|
| 1913 |
+
"step": 1355
|
| 1914 |
+
},
|
| 1915 |
+
{
|
| 1916 |
+
"epoch": 0.797686854754824,
|
| 1917 |
+
"grad_norm": 0.07373046875,
|
| 1918 |
+
"learning_rate": 2.383642998198731e-05,
|
| 1919 |
+
"loss": 1.066,
|
| 1920 |
+
"step": 1360
|
| 1921 |
+
},
|
| 1922 |
+
{
|
| 1923 |
+
"epoch": 0.8006195270149521,
|
| 1924 |
+
"grad_norm": 0.0732421875,
|
| 1925 |
+
"learning_rate": 2.317645772995042e-05,
|
| 1926 |
+
"loss": 1.0752,
|
| 1927 |
+
"step": 1365
|
| 1928 |
+
},
|
| 1929 |
+
{
|
| 1930 |
+
"epoch": 0.80355219927508,
|
| 1931 |
+
"grad_norm": 0.0751953125,
|
| 1932 |
+
"learning_rate": 2.2524551250497316e-05,
|
| 1933 |
+
"loss": 1.0738,
|
| 1934 |
+
"step": 1370
|
| 1935 |
+
},
|
| 1936 |
+
{
|
| 1937 |
+
"epoch": 0.8064848715352081,
|
| 1938 |
+
"grad_norm": 0.07373046875,
|
| 1939 |
+
"learning_rate": 2.1880778987873807e-05,
|
| 1940 |
+
"loss": 1.0717,
|
| 1941 |
+
"step": 1375
|
| 1942 |
+
},
|
| 1943 |
+
{
|
| 1944 |
+
"epoch": 0.8094175437953361,
|
| 1945 |
+
"grad_norm": 0.07470703125,
|
| 1946 |
+
"learning_rate": 2.124520853230697e-05,
|
| 1947 |
+
"loss": 1.0781,
|
| 1948 |
+
"step": 1380
|
| 1949 |
+
},
|
| 1950 |
+
{
|
| 1951 |
+
"epoch": 0.8123502160554642,
|
| 1952 |
+
"grad_norm": 0.07666015625,
|
| 1953 |
+
"learning_rate": 2.061790661290881e-05,
|
| 1954 |
+
"loss": 1.0774,
|
| 1955 |
+
"step": 1385
|
| 1956 |
+
},
|
| 1957 |
+
{
|
| 1958 |
+
"epoch": 0.8152828883155921,
|
| 1959 |
+
"grad_norm": 0.07568359375,
|
| 1960 |
+
"learning_rate": 1.999893909067021e-05,
|
| 1961 |
+
"loss": 1.0817,
|
| 1962 |
+
"step": 1390
|
| 1963 |
+
},
|
| 1964 |
+
{
|
| 1965 |
+
"epoch": 0.8182155605757202,
|
| 1966 |
+
"grad_norm": 0.0771484375,
|
| 1967 |
+
"learning_rate": 1.9388370951546432e-05,
|
| 1968 |
+
"loss": 1.0899,
|
| 1969 |
+
"step": 1395
|
| 1970 |
+
},
|
| 1971 |
+
{
|
| 1972 |
+
"epoch": 0.8211482328358483,
|
| 1973 |
+
"grad_norm": 0.076171875,
|
| 1974 |
+
"learning_rate": 1.8786266299633738e-05,
|
| 1975 |
+
"loss": 1.0724,
|
| 1976 |
+
"step": 1400
|
| 1977 |
+
},
|
| 1978 |
+
{
|
| 1979 |
+
"epoch": 0.8240809050959763,
|
| 1980 |
+
"grad_norm": 0.07421875,
|
| 1981 |
+
"learning_rate": 1.8192688350439424e-05,
|
| 1982 |
+
"loss": 1.0676,
|
| 1983 |
+
"step": 1405
|
| 1984 |
+
},
|
| 1985 |
+
{
|
| 1986 |
+
"epoch": 0.8270135773561043,
|
| 1987 |
+
"grad_norm": 0.0732421875,
|
| 1988 |
+
"learning_rate": 1.7607699424244585e-05,
|
| 1989 |
+
"loss": 1.0751,
|
| 1990 |
+
"step": 1410
|
| 1991 |
+
},
|
| 1992 |
+
{
|
| 1993 |
+
"epoch": 0.8299462496162323,
|
| 1994 |
+
"grad_norm": 0.07373046875,
|
| 1995 |
+
"learning_rate": 1.7031360939561103e-05,
|
| 1996 |
+
"loss": 1.0992,
|
| 1997 |
+
"step": 1415
|
| 1998 |
+
},
|
| 1999 |
+
{
|
| 2000 |
+
"epoch": 0.8328789218763604,
|
| 2001 |
+
"grad_norm": 0.07421875,
|
| 2002 |
+
"learning_rate": 1.646373340668319e-05,
|
| 2003 |
+
"loss": 1.0735,
|
| 2004 |
+
"step": 1420
|
| 2005 |
+
},
|
| 2006 |
+
{
|
| 2007 |
+
"epoch": 0.8358115941364884,
|
| 2008 |
+
"grad_norm": 0.0751953125,
|
| 2009 |
+
"learning_rate": 1.5904876421334536e-05,
|
| 2010 |
+
"loss": 1.0686,
|
| 2011 |
+
"step": 1425
|
| 2012 |
+
},
|
| 2013 |
+
{
|
| 2014 |
+
"epoch": 0.8387442663966165,
|
| 2015 |
+
"grad_norm": 0.07861328125,
|
| 2016 |
+
"learning_rate": 1.5354848658411048e-05,
|
| 2017 |
+
"loss": 1.079,
|
| 2018 |
+
"step": 1430
|
| 2019 |
+
},
|
| 2020 |
+
{
|
| 2021 |
+
"epoch": 0.8416769386567444,
|
| 2022 |
+
"grad_norm": 0.07470703125,
|
| 2023 |
+
"learning_rate": 1.4813707865820747e-05,
|
| 2024 |
+
"loss": 1.1048,
|
| 2025 |
+
"step": 1435
|
| 2026 |
+
},
|
| 2027 |
+
{
|
| 2028 |
+
"epoch": 0.8446096109168725,
|
| 2029 |
+
"grad_norm": 0.0751953125,
|
| 2030 |
+
"learning_rate": 1.4281510858420632e-05,
|
| 2031 |
+
"loss": 1.0656,
|
| 2032 |
+
"step": 1440
|
| 2033 |
+
},
|
| 2034 |
+
{
|
| 2035 |
+
"epoch": 0.8475422831770005,
|
| 2036 |
+
"grad_norm": 0.07470703125,
|
| 2037 |
+
"learning_rate": 1.3758313512051702e-05,
|
| 2038 |
+
"loss": 1.067,
|
| 2039 |
+
"step": 1445
|
| 2040 |
+
},
|
| 2041 |
+
{
|
| 2042 |
+
"epoch": 0.8504749554371286,
|
| 2043 |
+
"grad_norm": 0.07470703125,
|
| 2044 |
+
"learning_rate": 1.3244170757672425e-05,
|
| 2045 |
+
"loss": 1.0804,
|
| 2046 |
+
"step": 1450
|
| 2047 |
+
},
|
| 2048 |
+
{
|
| 2049 |
+
"epoch": 0.8534076276972565,
|
| 2050 |
+
"grad_norm": 0.0751953125,
|
| 2051 |
+
"learning_rate": 1.2739136575591581e-05,
|
| 2052 |
+
"loss": 1.0829,
|
| 2053 |
+
"step": 1455
|
| 2054 |
+
},
|
| 2055 |
+
{
|
| 2056 |
+
"epoch": 0.8563402999573846,
|
| 2057 |
+
"grad_norm": 0.07421875,
|
| 2058 |
+
"learning_rate": 1.2243263989800768e-05,
|
| 2059 |
+
"loss": 1.0792,
|
| 2060 |
+
"step": 1460
|
| 2061 |
+
},
|
| 2062 |
+
{
|
| 2063 |
+
"epoch": 0.8592729722175126,
|
| 2064 |
+
"grad_norm": 0.0751953125,
|
| 2065 |
+
"learning_rate": 1.17566050624074e-05,
|
| 2066 |
+
"loss": 1.0776,
|
| 2067 |
+
"step": 1465
|
| 2068 |
+
},
|
| 2069 |
+
{
|
| 2070 |
+
"epoch": 0.8622056444776407,
|
| 2071 |
+
"grad_norm": 0.0732421875,
|
| 2072 |
+
"learning_rate": 1.1279210888168546e-05,
|
| 2073 |
+
"loss": 1.0771,
|
| 2074 |
+
"step": 1470
|
| 2075 |
+
},
|
| 2076 |
+
{
|
| 2077 |
+
"epoch": 0.8651383167377686,
|
| 2078 |
+
"grad_norm": 0.0732421875,
|
| 2079 |
+
"learning_rate": 1.0811131589126667e-05,
|
| 2080 |
+
"loss": 1.0872,
|
| 2081 |
+
"step": 1475
|
| 2082 |
+
},
|
| 2083 |
+
{
|
| 2084 |
+
"epoch": 0.8680709889978967,
|
| 2085 |
+
"grad_norm": 0.0732421875,
|
| 2086 |
+
"learning_rate": 1.0352416309347001e-05,
|
| 2087 |
+
"loss": 1.0668,
|
| 2088 |
+
"step": 1480
|
| 2089 |
+
},
|
| 2090 |
+
{
|
| 2091 |
+
"epoch": 0.8710036612580248,
|
| 2092 |
+
"grad_norm": 0.07373046875,
|
| 2093 |
+
"learning_rate": 9.903113209758096e-06,
|
| 2094 |
+
"loss": 1.0598,
|
| 2095 |
+
"step": 1485
|
| 2096 |
+
},
|
| 2097 |
+
{
|
| 2098 |
+
"epoch": 0.8739363335181528,
|
| 2099 |
+
"grad_norm": 0.07275390625,
|
| 2100 |
+
"learning_rate": 9.463269463095203e-06,
|
| 2101 |
+
"loss": 1.0664,
|
| 2102 |
+
"step": 1490
|
| 2103 |
+
},
|
| 2104 |
+
{
|
| 2105 |
+
"epoch": 0.8768690057782809,
|
| 2106 |
+
"grad_norm": 0.07421875,
|
| 2107 |
+
"learning_rate": 9.032931248947685e-06,
|
| 2108 |
+
"loss": 1.0709,
|
| 2109 |
+
"step": 1495
|
| 2110 |
+
},
|
| 2111 |
+
{
|
| 2112 |
+
"epoch": 0.8798016780384088,
|
| 2113 |
+
"grad_norm": 0.0732421875,
|
| 2114 |
+
"learning_rate": 8.612143748910451e-06,
|
| 2115 |
+
"loss": 1.0585,
|
| 2116 |
+
"step": 1500
|
| 2117 |
+
},
|
| 2118 |
+
{
|
| 2119 |
+
"epoch": 0.8827343502985369,
|
| 2120 |
+
"grad_norm": 0.0771484375,
|
| 2121 |
+
"learning_rate": 8.20095114184044e-06,
|
| 2122 |
+
"loss": 1.0641,
|
| 2123 |
+
"step": 1505
|
| 2124 |
+
},
|
| 2125 |
+
{
|
| 2126 |
+
"epoch": 0.8856670225586649,
|
| 2127 |
+
"grad_norm": 0.07275390625,
|
| 2128 |
+
"learning_rate": 7.799396599218133e-06,
|
| 2129 |
+
"loss": 1.0674,
|
| 2130 |
+
"step": 1510
|
| 2131 |
+
},
|
| 2132 |
+
{
|
| 2133 |
+
"epoch": 0.888599694818793,
|
| 2134 |
+
"grad_norm": 0.07373046875,
|
| 2135 |
+
"learning_rate": 7.40752228061502e-06,
|
| 2136 |
+
"loss": 1.0844,
|
| 2137 |
+
"step": 1515
|
| 2138 |
+
},
|
| 2139 |
+
{
|
| 2140 |
+
"epoch": 0.8915323670789209,
|
| 2141 |
+
"grad_norm": 0.07373046875,
|
| 2142 |
+
"learning_rate": 7.0253693292671505e-06,
|
| 2143 |
+
"loss": 1.0832,
|
| 2144 |
+
"step": 1520
|
| 2145 |
+
},
|
| 2146 |
+
{
|
| 2147 |
+
"epoch": 0.894465039339049,
|
| 2148 |
+
"grad_norm": 0.07373046875,
|
| 2149 |
+
"learning_rate": 6.65297786775555e-06,
|
| 2150 |
+
"loss": 1.0834,
|
| 2151 |
+
"step": 1525
|
| 2152 |
+
},
|
| 2153 |
+
{
|
| 2154 |
+
"epoch": 0.897397711599177,
|
| 2155 |
+
"grad_norm": 0.0771484375,
|
| 2156 |
+
"learning_rate": 6.290386993793618e-06,
|
| 2157 |
+
"loss": 1.0561,
|
| 2158 |
+
"step": 1530
|
| 2159 |
+
},
|
| 2160 |
+
{
|
| 2161 |
+
"epoch": 0.9003303838593051,
|
| 2162 |
+
"grad_norm": 0.07373046875,
|
| 2163 |
+
"learning_rate": 5.937634776122348e-06,
|
| 2164 |
+
"loss": 1.0663,
|
| 2165 |
+
"step": 1535
|
| 2166 |
+
},
|
| 2167 |
+
{
|
| 2168 |
+
"epoch": 0.903263056119433,
|
| 2169 |
+
"grad_norm": 0.07666015625,
|
| 2170 |
+
"learning_rate": 5.594758250513333e-06,
|
| 2171 |
+
"loss": 1.0691,
|
| 2172 |
+
"step": 1540
|
| 2173 |
+
},
|
| 2174 |
+
{
|
| 2175 |
+
"epoch": 0.9061957283795611,
|
| 2176 |
+
"grad_norm": 0.07666015625,
|
| 2177 |
+
"learning_rate": 5.261793415880456e-06,
|
| 2178 |
+
"loss": 1.0672,
|
| 2179 |
+
"step": 1545
|
| 2180 |
+
},
|
| 2181 |
+
{
|
| 2182 |
+
"epoch": 0.9091284006396891,
|
| 2183 |
+
"grad_norm": 0.07568359375,
|
| 2184 |
+
"learning_rate": 4.938775230500192e-06,
|
| 2185 |
+
"loss": 1.0781,
|
| 2186 |
+
"step": 1550
|
| 2187 |
+
},
|
| 2188 |
+
{
|
| 2189 |
+
"epoch": 0.9120610728998172,
|
| 2190 |
+
"grad_norm": 0.07470703125,
|
| 2191 |
+
"learning_rate": 4.625737608341507e-06,
|
| 2192 |
+
"loss": 1.0764,
|
| 2193 |
+
"step": 1555
|
| 2194 |
+
},
|
| 2195 |
+
{
|
| 2196 |
+
"epoch": 0.9149937451599452,
|
| 2197 |
+
"grad_norm": 0.07421875,
|
| 2198 |
+
"learning_rate": 4.322713415504975e-06,
|
| 2199 |
+
"loss": 1.0858,
|
| 2200 |
+
"step": 1560
|
| 2201 |
+
},
|
| 2202 |
+
{
|
| 2203 |
+
"epoch": 0.9179264174200732,
|
| 2204 |
+
"grad_norm": 0.0732421875,
|
| 2205 |
+
"learning_rate": 4.029734466772328e-06,
|
| 2206 |
+
"loss": 1.0722,
|
| 2207 |
+
"step": 1565
|
| 2208 |
+
},
|
| 2209 |
+
{
|
| 2210 |
+
"epoch": 0.9208590896802012,
|
| 2211 |
+
"grad_norm": 0.07421875,
|
| 2212 |
+
"learning_rate": 3.7468315222660586e-06,
|
| 2213 |
+
"loss": 1.0667,
|
| 2214 |
+
"step": 1570
|
| 2215 |
+
},
|
| 2216 |
+
{
|
| 2217 |
+
"epoch": 0.9237917619403293,
|
| 2218 |
+
"grad_norm": 0.07421875,
|
| 2219 |
+
"learning_rate": 3.4740342842199956e-06,
|
| 2220 |
+
"loss": 1.0658,
|
| 2221 |
+
"step": 1575
|
| 2222 |
+
},
|
| 2223 |
+
{
|
| 2224 |
+
"epoch": 0.9267244342004574,
|
| 2225 |
+
"grad_norm": 0.07421875,
|
| 2226 |
+
"learning_rate": 3.211371393860718e-06,
|
| 2227 |
+
"loss": 1.0764,
|
| 2228 |
+
"step": 1580
|
| 2229 |
+
},
|
| 2230 |
+
{
|
| 2231 |
+
"epoch": 0.9296571064605853,
|
| 2232 |
+
"grad_norm": 0.07470703125,
|
| 2233 |
+
"learning_rate": 2.9588704284006174e-06,
|
| 2234 |
+
"loss": 1.0804,
|
| 2235 |
+
"step": 1585
|
| 2236 |
+
},
|
| 2237 |
+
{
|
| 2238 |
+
"epoch": 0.9325897787207134,
|
| 2239 |
+
"grad_norm": 0.0732421875,
|
| 2240 |
+
"learning_rate": 2.7165578981424357e-06,
|
| 2241 |
+
"loss": 1.0749,
|
| 2242 |
+
"step": 1590
|
| 2243 |
+
},
|
| 2244 |
+
{
|
| 2245 |
+
"epoch": 0.9355224509808414,
|
| 2246 |
+
"grad_norm": 0.07421875,
|
| 2247 |
+
"learning_rate": 2.484459243695991e-06,
|
| 2248 |
+
"loss": 1.0805,
|
| 2249 |
+
"step": 1595
|
| 2250 |
+
},
|
| 2251 |
+
{
|
| 2252 |
+
"epoch": 0.9384551232409695,
|
| 2253 |
+
"grad_norm": 0.07568359375,
|
| 2254 |
+
"learning_rate": 2.262598833307128e-06,
|
| 2255 |
+
"loss": 1.07,
|
| 2256 |
+
"step": 1600
|
| 2257 |
+
},
|
| 2258 |
+
{
|
| 2259 |
+
"epoch": 0.9413877955010974,
|
| 2260 |
+
"grad_norm": 0.07470703125,
|
| 2261 |
+
"learning_rate": 2.0509999602992493e-06,
|
| 2262 |
+
"loss": 1.0834,
|
| 2263 |
+
"step": 1605
|
| 2264 |
+
},
|
| 2265 |
+
{
|
| 2266 |
+
"epoch": 0.9443204677612255,
|
| 2267 |
+
"grad_norm": 0.0751953125,
|
| 2268 |
+
"learning_rate": 1.849684840627741e-06,
|
| 2269 |
+
"loss": 1.0646,
|
| 2270 |
+
"step": 1610
|
| 2271 |
+
},
|
| 2272 |
+
{
|
| 2273 |
+
"epoch": 0.9472531400213535,
|
| 2274 |
+
"grad_norm": 0.07373046875,
|
| 2275 |
+
"learning_rate": 1.6586746105475281e-06,
|
| 2276 |
+
"loss": 1.0626,
|
| 2277 |
+
"step": 1615
|
| 2278 |
+
},
|
| 2279 |
+
{
|
| 2280 |
+
"epoch": 0.9501858122814816,
|
| 2281 |
+
"grad_norm": 0.0712890625,
|
| 2282 |
+
"learning_rate": 1.4779893243939359e-06,
|
| 2283 |
+
"loss": 1.0706,
|
| 2284 |
+
"step": 1620
|
| 2285 |
+
},
|
| 2286 |
+
{
|
| 2287 |
+
"epoch": 0.9531184845416096,
|
| 2288 |
+
"grad_norm": 0.076171875,
|
| 2289 |
+
"learning_rate": 1.3076479524771644e-06,
|
| 2290 |
+
"loss": 1.0747,
|
| 2291 |
+
"step": 1625
|
| 2292 |
+
},
|
| 2293 |
+
{
|
| 2294 |
+
"epoch": 0.9560511568017376,
|
| 2295 |
+
"grad_norm": 0.07470703125,
|
| 2296 |
+
"learning_rate": 1.1476683790905495e-06,
|
| 2297 |
+
"loss": 1.0859,
|
| 2298 |
+
"step": 1630
|
| 2299 |
+
},
|
| 2300 |
+
{
|
| 2301 |
+
"epoch": 0.9589838290618656,
|
| 2302 |
+
"grad_norm": 0.07421875,
|
| 2303 |
+
"learning_rate": 9.98067400632985e-07,
|
| 2304 |
+
"loss": 1.0916,
|
| 2305 |
+
"step": 1635
|
| 2306 |
+
},
|
| 2307 |
+
{
|
| 2308 |
+
"epoch": 0.9619165013219937,
|
| 2309 |
+
"grad_norm": 0.0751953125,
|
| 2310 |
+
"learning_rate": 8.588607238453006e-07,
|
| 2311 |
+
"loss": 1.0644,
|
| 2312 |
+
"step": 1640
|
| 2313 |
+
},
|
| 2314 |
+
{
|
| 2315 |
+
"epoch": 0.9648491735821217,
|
| 2316 |
+
"grad_norm": 0.07421875,
|
| 2317 |
+
"learning_rate": 7.300629641613154e-07,
|
| 2318 |
+
"loss": 1.0783,
|
| 2319 |
+
"step": 1645
|
| 2320 |
+
},
|
| 2321 |
+
{
|
| 2322 |
+
"epoch": 0.9677818458422497,
|
| 2323 |
+
"grad_norm": 0.07421875,
|
| 2324 |
+
"learning_rate": 6.116876441733088e-07,
|
| 2325 |
+
"loss": 1.0885,
|
| 2326 |
+
"step": 1650
|
| 2327 |
+
},
|
| 2328 |
+
{
|
| 2329 |
+
"epoch": 0.9707145181023777,
|
| 2330 |
+
"grad_norm": 0.07373046875,
|
| 2331 |
+
"learning_rate": 5.037471922122561e-07,
|
| 2332 |
+
"loss": 1.0759,
|
| 2333 |
+
"step": 1655
|
| 2334 |
+
},
|
| 2335 |
+
{
|
| 2336 |
+
"epoch": 0.9736471903625058,
|
| 2337 |
+
"grad_norm": 0.07470703125,
|
| 2338 |
+
"learning_rate": 4.062529410429949e-07,
|
| 2339 |
+
"loss": 1.0633,
|
| 2340 |
+
"step": 1660
|
| 2341 |
+
},
|
| 2342 |
+
{
|
| 2343 |
+
"epoch": 0.9765798626226339,
|
| 2344 |
+
"grad_norm": 0.07373046875,
|
| 2345 |
+
"learning_rate": 3.192151266743548e-07,
|
| 2346 |
+
"loss": 1.0794,
|
| 2347 |
+
"step": 1665
|
| 2348 |
+
},
|
| 2349 |
+
{
|
| 2350 |
+
"epoch": 0.9795125348827618,
|
| 2351 |
+
"grad_norm": 0.072265625,
|
| 2352 |
+
"learning_rate": 2.4264288728451657e-07,
|
| 2353 |
+
"loss": 1.0704,
|
| 2354 |
+
"step": 1670
|
| 2355 |
+
},
|
| 2356 |
+
{
|
| 2357 |
+
"epoch": 0.9824452071428899,
|
| 2358 |
+
"grad_norm": 0.0751953125,
|
| 2359 |
+
"learning_rate": 1.7654426226155763e-07,
|
| 2360 |
+
"loss": 1.0678,
|
| 2361 |
+
"step": 1675
|
| 2362 |
+
},
|
| 2363 |
+
{
|
| 2364 |
+
"epoch": 0.9853778794030179,
|
| 2365 |
+
"grad_norm": 0.07275390625,
|
| 2366 |
+
"learning_rate": 1.2092619135937177e-07,
|
| 2367 |
+
"loss": 1.067,
|
| 2368 |
+
"step": 1680
|
| 2369 |
+
},
|
| 2370 |
+
{
|
| 2371 |
+
"epoch": 0.988310551663146,
|
| 2372 |
+
"grad_norm": 0.07373046875,
|
| 2373 |
+
"learning_rate": 7.579451396908521e-08,
|
| 2374 |
+
"loss": 1.0654,
|
| 2375 |
+
"step": 1685
|
| 2376 |
+
},
|
| 2377 |
+
{
|
| 2378 |
+
"epoch": 0.9912432239232739,
|
| 2379 |
+
"grad_norm": 0.0751953125,
|
| 2380 |
+
"learning_rate": 4.1153968505991406e-08,
|
| 2381 |
+
"loss": 1.0512,
|
| 2382 |
+
"step": 1690
|
| 2383 |
+
},
|
| 2384 |
+
{
|
| 2385 |
+
"epoch": 0.994175896183402,
|
| 2386 |
+
"grad_norm": 0.07568359375,
|
| 2387 |
+
"learning_rate": 1.7008191912004646e-08,
|
| 2388 |
+
"loss": 1.075,
|
| 2389 |
+
"step": 1695
|
| 2390 |
+
},
|
| 2391 |
+
{
|
| 2392 |
+
"epoch": 0.99710856844353,
|
| 2393 |
+
"grad_norm": 0.07568359375,
|
| 2394 |
+
"learning_rate": 3.359719273865469e-09,
|
| 2395 |
+
"loss": 1.0838,
|
| 2396 |
+
"step": 1700
|
| 2397 |
+
},
|
| 2398 |
+
{
|
| 2399 |
+
"epoch": 0.9994547062516325,
|
| 2400 |
+
"eval_loss": 1.1863435506820679,
|
| 2401 |
+
"eval_runtime": 1954.7038,
|
| 2402 |
+
"eval_samples_per_second": 8.263,
|
| 2403 |
+
"eval_steps_per_second": 8.263,
|
| 2404 |
+
"step": 1704
|
| 2405 |
+
},
|
| 2406 |
+
{
|
| 2407 |
+
"epoch": 0.9994547062516325,
|
| 2408 |
+
"step": 1704,
|
| 2409 |
+
"total_flos": 4.734725057610252e+18,
|
| 2410 |
+
"train_loss": 0.12850537406446788,
|
| 2411 |
+
"train_runtime": 13780.2764,
|
| 2412 |
+
"train_samples_per_second": 15.836,
|
| 2413 |
+
"train_steps_per_second": 0.124
|
| 2414 |
+
}
|
| 2415 |
+
],
|
| 2416 |
+
"logging_steps": 5,
|
| 2417 |
+
"max_steps": 1704,
|
| 2418 |
+
"num_input_tokens_seen": 0,
|
| 2419 |
+
"num_train_epochs": 1,
|
| 2420 |
+
"save_steps": 20,
|
| 2421 |
+
"total_flos": 4.734725057610252e+18,
|
| 2422 |
+
"train_batch_size": 1,
|
| 2423 |
+
"trial_name": null,
|
| 2424 |
+
"trial_params": null
|
| 2425 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:bf76d159073419817d3d6a48d28e401dc4df692bd0ef04b9aff382938e6ccbb7
|
| 3 |
+
size 5112
|