Upload folder using huggingface_hub
Browse files- README.md +57 -0
- all_results.json +12 -0
- checkpoint-2940/config.json +32 -0
- checkpoint-2940/generation_config.json +9 -0
- checkpoint-2940/model.safetensors +3 -0
- checkpoint-2940/optimizer.pt +3 -0
- checkpoint-2940/rng_state.pth +3 -0
- checkpoint-2940/scheduler.pt +3 -0
- checkpoint-2940/tokenizer.json +0 -0
- checkpoint-2940/tokenizer_config.json +21 -0
- checkpoint-2940/trainer_state.json +669 -0
- checkpoint-2940/training_args.bin +3 -0
- checkpoint-5880/config.json +32 -0
- checkpoint-5880/generation_config.json +9 -0
- checkpoint-5880/model.safetensors +3 -0
- checkpoint-5880/optimizer.pt +3 -0
- checkpoint-5880/rng_state.pth +3 -0
- checkpoint-5880/scheduler.pt +3 -0
- checkpoint-5880/tokenizer.json +0 -0
- checkpoint-5880/tokenizer_config.json +21 -0
- checkpoint-5880/trainer_state.json +1325 -0
- checkpoint-5880/training_args.bin +3 -0
- checkpoint-8820/config.json +32 -0
- checkpoint-8820/generation_config.json +9 -0
- checkpoint-8820/model.safetensors +3 -0
- checkpoint-8820/optimizer.pt +3 -0
- checkpoint-8820/rng_state.pth +3 -0
- checkpoint-8820/scheduler.pt +3 -0
- checkpoint-8820/tokenizer.json +0 -0
- checkpoint-8820/tokenizer_config.json +21 -0
- checkpoint-8820/trainer_state.json +1981 -0
- checkpoint-8820/training_args.bin +3 -0
- eval_metrics.json +6 -0
- eval_results.json +6 -0
- final/config.json +32 -0
- final/generation_config.json +9 -0
- final/model.safetensors +3 -0
- final/tokenizer.json +0 -0
- final/tokenizer_config.json +21 -0
- final/training_args.bin +3 -0
- train_metrics.json +8 -0
- train_results.json +8 -0
- trainer_state.json +1993 -0
README.md
ADDED
|
@@ -0,0 +1,57 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
library_name: transformers
|
| 3 |
+
model_name: sft-trellis
|
| 4 |
+
tags:
|
| 5 |
+
- generated_from_trainer
|
| 6 |
+
- sft
|
| 7 |
+
- trl
|
| 8 |
+
licence: license
|
| 9 |
+
---
|
| 10 |
+
|
| 11 |
+
# Model Card for sft-trellis
|
| 12 |
+
|
| 13 |
+
This model is a fine-tuned version of [None](https://huggingface.co/None).
|
| 14 |
+
It has been trained using [TRL](https://github.com/huggingface/trl).
|
| 15 |
+
|
| 16 |
+
## Quick start
|
| 17 |
+
|
| 18 |
+
```python
|
| 19 |
+
from transformers import pipeline
|
| 20 |
+
|
| 21 |
+
question = "If you had a time machine, but could only go to the past or the future once and never return, which would you choose and why?"
|
| 22 |
+
generator = pipeline("text-generation", model="None", device="cuda")
|
| 23 |
+
output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0]
|
| 24 |
+
print(output["generated_text"])
|
| 25 |
+
```
|
| 26 |
+
|
| 27 |
+
## Training procedure
|
| 28 |
+
|
| 29 |
+
|
| 30 |
+
|
| 31 |
+
|
| 32 |
+
|
| 33 |
+
This model was trained with SFT.
|
| 34 |
+
|
| 35 |
+
### Framework versions
|
| 36 |
+
|
| 37 |
+
- TRL: 0.29.0
|
| 38 |
+
- Transformers: 5.3.0
|
| 39 |
+
- Pytorch: 2.8.0+cu128
|
| 40 |
+
- Datasets: 4.7.0
|
| 41 |
+
- Tokenizers: 0.22.2
|
| 42 |
+
|
| 43 |
+
## Citations
|
| 44 |
+
|
| 45 |
+
|
| 46 |
+
|
| 47 |
+
Cite TRL as:
|
| 48 |
+
|
| 49 |
+
```bibtex
|
| 50 |
+
@software{vonwerra2020trl,
|
| 51 |
+
title = {{TRL: Transformers Reinforcement Learning}},
|
| 52 |
+
author = {von Werra, Leandro and Belkada, Younes and Tunstall, Lewis and Beeching, Edward and Thrush, Tristan and Lambert, Nathan and Huang, Shengyi and Rasul, Kashif and Gallouédec, Quentin},
|
| 53 |
+
license = {Apache-2.0},
|
| 54 |
+
url = {https://github.com/huggingface/trl},
|
| 55 |
+
year = {2020}
|
| 56 |
+
}
|
| 57 |
+
```
|
all_results.json
ADDED
|
@@ -0,0 +1,12 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"eval_loss": 1.1048007011413574,
|
| 3 |
+
"eval_runtime": 8.5485,
|
| 4 |
+
"eval_samples_per_second": 110.897,
|
| 5 |
+
"eval_steps_per_second": 13.921,
|
| 6 |
+
"total_flos": 1.3775254383370368e+18,
|
| 7 |
+
"train_examples": 94067,
|
| 8 |
+
"train_loss": 1.158577897164827,
|
| 9 |
+
"train_runtime": 2757.2959,
|
| 10 |
+
"train_samples_per_second": 102.347,
|
| 11 |
+
"train_steps_per_second": 3.199
|
| 12 |
+
}
|
checkpoint-2940/config.json
ADDED
|
@@ -0,0 +1,32 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"LlamaForCausalLM"
|
| 4 |
+
],
|
| 5 |
+
"attention_bias": false,
|
| 6 |
+
"attention_dropout": 0.0,
|
| 7 |
+
"bos_token_id": 0,
|
| 8 |
+
"dtype": "bfloat16",
|
| 9 |
+
"eos_token_id": 0,
|
| 10 |
+
"head_dim": 64,
|
| 11 |
+
"hidden_act": "silu",
|
| 12 |
+
"hidden_size": 1280,
|
| 13 |
+
"initializer_range": 0.02,
|
| 14 |
+
"intermediate_size": 2816,
|
| 15 |
+
"max_position_embeddings": 2048,
|
| 16 |
+
"mlp_bias": false,
|
| 17 |
+
"model_type": "llama",
|
| 18 |
+
"num_attention_heads": 20,
|
| 19 |
+
"num_hidden_layers": 24,
|
| 20 |
+
"num_key_value_heads": 10,
|
| 21 |
+
"pad_token_id": 1,
|
| 22 |
+
"pretraining_tp": 1,
|
| 23 |
+
"rms_norm_eps": 1e-05,
|
| 24 |
+
"rope_parameters": {
|
| 25 |
+
"rope_theta": 10000.0,
|
| 26 |
+
"rope_type": "default"
|
| 27 |
+
},
|
| 28 |
+
"tie_word_embeddings": false,
|
| 29 |
+
"transformers_version": "5.3.0",
|
| 30 |
+
"use_cache": false,
|
| 31 |
+
"vocab_size": 50304
|
| 32 |
+
}
|
checkpoint-2940/generation_config.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"bos_token_id": 0,
|
| 4 |
+
"eos_token_id": [
|
| 5 |
+
0
|
| 6 |
+
],
|
| 7 |
+
"pad_token_id": 1,
|
| 8 |
+
"transformers_version": "5.3.0"
|
| 9 |
+
}
|
checkpoint-2940/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8e5669fd10d44d8360010504e183ab1530bc921a4898742d28d668f97c0f0a5e
|
| 3 |
+
size 1012681552
|
checkpoint-2940/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3204898e3f783707ae155f46ac88dca5f0f63285c0625bdca9f46ea4919105cd
|
| 3 |
+
size 2025501387
|
checkpoint-2940/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:01f9a0f7843a37be87edd23f4e88aa93b38b95cc2c07503eeb1cf2e4632453a2
|
| 3 |
+
size 14645
|
checkpoint-2940/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:54af7fda372ac423805aa02a135480c56b80c94ab808518ea36ef152fe12cc92
|
| 3 |
+
size 1465
|
checkpoint-2940/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
checkpoint-2940/tokenizer_config.json
ADDED
|
@@ -0,0 +1,21 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_prefix_space": false,
|
| 3 |
+
"backend": "tokenizers",
|
| 4 |
+
"bos_token": "<|endoftext|>",
|
| 5 |
+
"eos_token": "<|endoftext|>",
|
| 6 |
+
"errors": "replace",
|
| 7 |
+
"extra_special_tokens": [
|
| 8 |
+
"<|system|>",
|
| 9 |
+
"<|user|>",
|
| 10 |
+
"<|assistant|>",
|
| 11 |
+
"<|tool_call|>",
|
| 12 |
+
"<|tool_result|>",
|
| 13 |
+
"<|end|>"
|
| 14 |
+
],
|
| 15 |
+
"is_local": true,
|
| 16 |
+
"model_max_length": 1000000000000000019884624838656,
|
| 17 |
+
"pad_token": "<|padding|>",
|
| 18 |
+
"tokenizer_class": "GPTNeoXTokenizer",
|
| 19 |
+
"trim_offsets": true,
|
| 20 |
+
"unk_token": "<|endoftext|>"
|
| 21 |
+
}
|
checkpoint-2940/trainer_state.json
ADDED
|
@@ -0,0 +1,669 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": null,
|
| 3 |
+
"best_metric": null,
|
| 4 |
+
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 1.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 2940,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"entropy": 1.7698939561843872,
|
| 14 |
+
"epoch": 0.017006802721088437,
|
| 15 |
+
"grad_norm": 3.796875,
|
| 16 |
+
"learning_rate": 1.112372304199773e-06,
|
| 17 |
+
"loss": 1.5452156066894531,
|
| 18 |
+
"mean_token_accuracy": 0.6630511581897736,
|
| 19 |
+
"num_tokens": 1416855.0,
|
| 20 |
+
"step": 50
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"entropy": 1.765206663608551,
|
| 24 |
+
"epoch": 0.034013605442176874,
|
| 25 |
+
"grad_norm": 4.6875,
|
| 26 |
+
"learning_rate": 2.2474460839954596e-06,
|
| 27 |
+
"loss": 1.5233023071289062,
|
| 28 |
+
"mean_token_accuracy": 0.6625268197059632,
|
| 29 |
+
"num_tokens": 2861955.0,
|
| 30 |
+
"step": 100
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"entropy": 1.7640055298805237,
|
| 34 |
+
"epoch": 0.05102040816326531,
|
| 35 |
+
"grad_norm": 2.984375,
|
| 36 |
+
"learning_rate": 3.3825198637911465e-06,
|
| 37 |
+
"loss": 1.4435250854492188,
|
| 38 |
+
"mean_token_accuracy": 0.6652392935752869,
|
| 39 |
+
"num_tokens": 4279472.0,
|
| 40 |
+
"step": 150
|
| 41 |
+
},
|
| 42 |
+
{
|
| 43 |
+
"entropy": 1.7068938755989074,
|
| 44 |
+
"epoch": 0.06802721088435375,
|
| 45 |
+
"grad_norm": 1.3828125,
|
| 46 |
+
"learning_rate": 4.517593643586833e-06,
|
| 47 |
+
"loss": 1.3378631591796875,
|
| 48 |
+
"mean_token_accuracy": 0.673524329662323,
|
| 49 |
+
"num_tokens": 5707775.0,
|
| 50 |
+
"step": 200
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"entropy": 1.6458502221107483,
|
| 54 |
+
"epoch": 0.08503401360544217,
|
| 55 |
+
"grad_norm": 1.1015625,
|
| 56 |
+
"learning_rate": 5.652667423382521e-06,
|
| 57 |
+
"loss": 1.2989051818847657,
|
| 58 |
+
"mean_token_accuracy": 0.6784149742126465,
|
| 59 |
+
"num_tokens": 7118906.0,
|
| 60 |
+
"step": 250
|
| 61 |
+
},
|
| 62 |
+
{
|
| 63 |
+
"entropy": 1.6340991377830505,
|
| 64 |
+
"epoch": 0.10204081632653061,
|
| 65 |
+
"grad_norm": 1.21875,
|
| 66 |
+
"learning_rate": 6.787741203178207e-06,
|
| 67 |
+
"loss": 1.2801470947265625,
|
| 68 |
+
"mean_token_accuracy": 0.6825140416622162,
|
| 69 |
+
"num_tokens": 8529671.0,
|
| 70 |
+
"step": 300
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"entropy": 1.599980125427246,
|
| 74 |
+
"epoch": 0.11904761904761904,
|
| 75 |
+
"grad_norm": 1.1640625,
|
| 76 |
+
"learning_rate": 7.922814982973895e-06,
|
| 77 |
+
"loss": 1.2615231323242186,
|
| 78 |
+
"mean_token_accuracy": 0.6857551598548889,
|
| 79 |
+
"num_tokens": 9973401.0,
|
| 80 |
+
"step": 350
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"entropy": 1.570529818534851,
|
| 84 |
+
"epoch": 0.1360544217687075,
|
| 85 |
+
"grad_norm": 1.078125,
|
| 86 |
+
"learning_rate": 9.057888762769582e-06,
|
| 87 |
+
"loss": 1.2502473449707032,
|
| 88 |
+
"mean_token_accuracy": 0.6881939315795899,
|
| 89 |
+
"num_tokens": 11391611.0,
|
| 90 |
+
"step": 400
|
| 91 |
+
},
|
| 92 |
+
{
|
| 93 |
+
"entropy": 1.5739640021324157,
|
| 94 |
+
"epoch": 0.15306122448979592,
|
| 95 |
+
"grad_norm": 1.1640625,
|
| 96 |
+
"learning_rate": 1.0192962542565269e-05,
|
| 97 |
+
"loss": 1.2381304168701173,
|
| 98 |
+
"mean_token_accuracy": 0.6904172372817993,
|
| 99 |
+
"num_tokens": 12818742.0,
|
| 100 |
+
"step": 450
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"entropy": 1.542729480266571,
|
| 104 |
+
"epoch": 0.17006802721088435,
|
| 105 |
+
"grad_norm": 1.2421875,
|
| 106 |
+
"learning_rate": 1.1328036322360954e-05,
|
| 107 |
+
"loss": 1.2234064483642577,
|
| 108 |
+
"mean_token_accuracy": 0.6928502190113067,
|
| 109 |
+
"num_tokens": 14230694.0,
|
| 110 |
+
"step": 500
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.17006802721088435,
|
| 114 |
+
"eval_entropy": 1.5409097210699771,
|
| 115 |
+
"eval_loss": 1.1968441009521484,
|
| 116 |
+
"eval_mean_token_accuracy": 0.6989319805337602,
|
| 117 |
+
"eval_num_tokens": 14230694.0,
|
| 118 |
+
"eval_runtime": 9.3471,
|
| 119 |
+
"eval_samples_per_second": 101.422,
|
| 120 |
+
"eval_steps_per_second": 12.731,
|
| 121 |
+
"step": 500
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"entropy": 1.5507296323776245,
|
| 125 |
+
"epoch": 0.1870748299319728,
|
| 126 |
+
"grad_norm": 1.2265625,
|
| 127 |
+
"learning_rate": 1.246311010215664e-05,
|
| 128 |
+
"loss": 1.2286612701416015,
|
| 129 |
+
"mean_token_accuracy": 0.6924433386325837,
|
| 130 |
+
"num_tokens": 15656461.0,
|
| 131 |
+
"step": 550
|
| 132 |
+
},
|
| 133 |
+
{
|
| 134 |
+
"entropy": 1.5728167843818666,
|
| 135 |
+
"epoch": 0.20408163265306123,
|
| 136 |
+
"grad_norm": 1.21875,
|
| 137 |
+
"learning_rate": 1.3598183881952328e-05,
|
| 138 |
+
"loss": 1.2478004455566407,
|
| 139 |
+
"mean_token_accuracy": 0.6871815168857575,
|
| 140 |
+
"num_tokens": 17079797.0,
|
| 141 |
+
"step": 600
|
| 142 |
+
},
|
| 143 |
+
{
|
| 144 |
+
"entropy": 1.5617320823669434,
|
| 145 |
+
"epoch": 0.22108843537414966,
|
| 146 |
+
"grad_norm": 1.125,
|
| 147 |
+
"learning_rate": 1.4733257661748014e-05,
|
| 148 |
+
"loss": 1.2422699737548828,
|
| 149 |
+
"mean_token_accuracy": 0.6885153651237488,
|
| 150 |
+
"num_tokens": 18519114.0,
|
| 151 |
+
"step": 650
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"entropy": 1.54343731880188,
|
| 155 |
+
"epoch": 0.23809523809523808,
|
| 156 |
+
"grad_norm": 1.1484375,
|
| 157 |
+
"learning_rate": 1.58683314415437e-05,
|
| 158 |
+
"loss": 1.2279574584960937,
|
| 159 |
+
"mean_token_accuracy": 0.6913824212551117,
|
| 160 |
+
"num_tokens": 19930551.0,
|
| 161 |
+
"step": 700
|
| 162 |
+
},
|
| 163 |
+
{
|
| 164 |
+
"entropy": 1.524279227256775,
|
| 165 |
+
"epoch": 0.25510204081632654,
|
| 166 |
+
"grad_norm": 1.140625,
|
| 167 |
+
"learning_rate": 1.700340522133939e-05,
|
| 168 |
+
"loss": 1.2193751525878906,
|
| 169 |
+
"mean_token_accuracy": 0.6942444491386414,
|
| 170 |
+
"num_tokens": 21345785.0,
|
| 171 |
+
"step": 750
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"entropy": 1.527976930141449,
|
| 175 |
+
"epoch": 0.272108843537415,
|
| 176 |
+
"grad_norm": 1.1171875,
|
| 177 |
+
"learning_rate": 1.8138479001135073e-05,
|
| 178 |
+
"loss": 1.212171859741211,
|
| 179 |
+
"mean_token_accuracy": 0.6960112357139587,
|
| 180 |
+
"num_tokens": 22754238.0,
|
| 181 |
+
"step": 800
|
| 182 |
+
},
|
| 183 |
+
{
|
| 184 |
+
"entropy": 1.5164598178863526,
|
| 185 |
+
"epoch": 0.2891156462585034,
|
| 186 |
+
"grad_norm": 1.109375,
|
| 187 |
+
"learning_rate": 1.9273552780930764e-05,
|
| 188 |
+
"loss": 1.2044546508789062,
|
| 189 |
+
"mean_token_accuracy": 0.6967348265647888,
|
| 190 |
+
"num_tokens": 24167833.0,
|
| 191 |
+
"step": 850
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"entropy": 1.4968056178092957,
|
| 195 |
+
"epoch": 0.30612244897959184,
|
| 196 |
+
"grad_norm": 1.0234375,
|
| 197 |
+
"learning_rate": 1.9999746322867523e-05,
|
| 198 |
+
"loss": 1.2030063629150392,
|
| 199 |
+
"mean_token_accuracy": 0.6967223298549652,
|
| 200 |
+
"num_tokens": 25597839.0,
|
| 201 |
+
"step": 900
|
| 202 |
+
},
|
| 203 |
+
{
|
| 204 |
+
"entropy": 1.5187969017028808,
|
| 205 |
+
"epoch": 0.3231292517006803,
|
| 206 |
+
"grad_norm": 1.1796875,
|
| 207 |
+
"learning_rate": 1.9996379823325586e-05,
|
| 208 |
+
"loss": 1.2039017486572265,
|
| 209 |
+
"mean_token_accuracy": 0.69672048330307,
|
| 210 |
+
"num_tokens": 27039491.0,
|
| 211 |
+
"step": 950
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"entropy": 1.5000163841247558,
|
| 215 |
+
"epoch": 0.3401360544217687,
|
| 216 |
+
"grad_norm": 1.0546875,
|
| 217 |
+
"learning_rate": 1.9989100081556328e-05,
|
| 218 |
+
"loss": 1.2096540832519531,
|
| 219 |
+
"mean_token_accuracy": 0.696036319732666,
|
| 220 |
+
"num_tokens": 28462525.0,
|
| 221 |
+
"step": 1000
|
| 222 |
+
},
|
| 223 |
+
{
|
| 224 |
+
"epoch": 0.3401360544217687,
|
| 225 |
+
"eval_entropy": 1.4825329389892707,
|
| 226 |
+
"eval_loss": 1.1529995203018188,
|
| 227 |
+
"eval_mean_token_accuracy": 0.7072621358542883,
|
| 228 |
+
"eval_num_tokens": 28462525.0,
|
| 229 |
+
"eval_runtime": 9.42,
|
| 230 |
+
"eval_samples_per_second": 100.637,
|
| 231 |
+
"eval_steps_per_second": 12.633,
|
| 232 |
+
"step": 1000
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"entropy": 1.4857492208480836,
|
| 236 |
+
"epoch": 0.35714285714285715,
|
| 237 |
+
"grad_norm": 1.0625,
|
| 238 |
+
"learning_rate": 1.997790994733071e-05,
|
| 239 |
+
"loss": 1.1972512054443358,
|
| 240 |
+
"mean_token_accuracy": 0.6976107740402222,
|
| 241 |
+
"num_tokens": 29880725.0,
|
| 242 |
+
"step": 1050
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"entropy": 1.4679273128509522,
|
| 246 |
+
"epoch": 0.3741496598639456,
|
| 247 |
+
"grad_norm": 1.125,
|
| 248 |
+
"learning_rate": 1.9962813801205148e-05,
|
| 249 |
+
"loss": 1.1646170806884766,
|
| 250 |
+
"mean_token_accuracy": 0.705207599401474,
|
| 251 |
+
"num_tokens": 31288250.0,
|
| 252 |
+
"step": 1100
|
| 253 |
+
},
|
| 254 |
+
{
|
| 255 |
+
"entropy": 1.482924530506134,
|
| 256 |
+
"epoch": 0.391156462585034,
|
| 257 |
+
"grad_norm": 1.25,
|
| 258 |
+
"learning_rate": 1.9943817552806678e-05,
|
| 259 |
+
"loss": 1.1940898132324218,
|
| 260 |
+
"mean_token_accuracy": 0.6988638079166413,
|
| 261 |
+
"num_tokens": 32696424.0,
|
| 262 |
+
"step": 1150
|
| 263 |
+
},
|
| 264 |
+
{
|
| 265 |
+
"entropy": 1.4768511247634888,
|
| 266 |
+
"epoch": 0.40816326530612246,
|
| 267 |
+
"grad_norm": 1.1015625,
|
| 268 |
+
"learning_rate": 1.9920928638519545e-05,
|
| 269 |
+
"loss": 1.1782675933837892,
|
| 270 |
+
"mean_token_accuracy": 0.7024707806110382,
|
| 271 |
+
"num_tokens": 34115752.0,
|
| 272 |
+
"step": 1200
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"entropy": 1.4849171042442322,
|
| 276 |
+
"epoch": 0.42517006802721086,
|
| 277 |
+
"grad_norm": 1.1015625,
|
| 278 |
+
"learning_rate": 1.9894156018574104e-05,
|
| 279 |
+
"loss": 1.1730673217773437,
|
| 280 |
+
"mean_token_accuracy": 0.7031274008750915,
|
| 281 |
+
"num_tokens": 35496076.0,
|
| 282 |
+
"step": 1250
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"entropy": 1.477535102367401,
|
| 286 |
+
"epoch": 0.4421768707482993,
|
| 287 |
+
"grad_norm": 1.140625,
|
| 288 |
+
"learning_rate": 1.9863510173539193e-05,
|
| 289 |
+
"loss": 1.1849075317382813,
|
| 290 |
+
"mean_token_accuracy": 0.6999362814426422,
|
| 291 |
+
"num_tokens": 36924227.0,
|
| 292 |
+
"step": 1300
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"entropy": 1.4699360299110413,
|
| 296 |
+
"epoch": 0.45918367346938777,
|
| 297 |
+
"grad_norm": 0.9921875,
|
| 298 |
+
"learning_rate": 1.982900310021936e-05,
|
| 299 |
+
"loss": 1.1758278656005858,
|
| 300 |
+
"mean_token_accuracy": 0.7016613483428955,
|
| 301 |
+
"num_tokens": 38298695.0,
|
| 302 |
+
"step": 1350
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"entropy": 1.4796976780891418,
|
| 306 |
+
"epoch": 0.47619047619047616,
|
| 307 |
+
"grad_norm": 1.2421875,
|
| 308 |
+
"learning_rate": 1.9790648306958504e-05,
|
| 309 |
+
"loss": 1.1764393615722657,
|
| 310 |
+
"mean_token_accuracy": 0.7006231868267059,
|
| 311 |
+
"num_tokens": 39730904.0,
|
| 312 |
+
"step": 1400
|
| 313 |
+
},
|
| 314 |
+
{
|
| 315 |
+
"entropy": 1.4659403300285339,
|
| 316 |
+
"epoch": 0.4931972789115646,
|
| 317 |
+
"grad_norm": 1.109375,
|
| 318 |
+
"learning_rate": 1.974846080835184e-05,
|
| 319 |
+
"loss": 1.1739279174804687,
|
| 320 |
+
"mean_token_accuracy": 0.7042246878147125,
|
| 321 |
+
"num_tokens": 41159605.0,
|
| 322 |
+
"step": 1450
|
| 323 |
+
},
|
| 324 |
+
{
|
| 325 |
+
"entropy": 1.4660640573501587,
|
| 326 |
+
"epoch": 0.5102040816326531,
|
| 327 |
+
"grad_norm": 1.15625,
|
| 328 |
+
"learning_rate": 1.9702457119368175e-05,
|
| 329 |
+
"loss": 1.1742501831054688,
|
| 330 |
+
"mean_token_accuracy": 0.7028728783130646,
|
| 331 |
+
"num_tokens": 42591337.0,
|
| 332 |
+
"step": 1500
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 0.5102040816326531,
|
| 336 |
+
"eval_entropy": 1.4500642824573677,
|
| 337 |
+
"eval_loss": 1.1341153383255005,
|
| 338 |
+
"eval_mean_token_accuracy": 0.7110618072397569,
|
| 339 |
+
"eval_num_tokens": 42591337.0,
|
| 340 |
+
"eval_runtime": 9.205,
|
| 341 |
+
"eval_samples_per_second": 102.988,
|
| 342 |
+
"eval_steps_per_second": 12.928,
|
| 343 |
+
"step": 1500
|
| 344 |
+
},
|
| 345 |
+
{
|
| 346 |
+
"entropy": 1.4817289400100708,
|
| 347 |
+
"epoch": 0.5272108843537415,
|
| 348 |
+
"grad_norm": 1.15625,
|
| 349 |
+
"learning_rate": 1.965265524888488e-05,
|
| 350 |
+
"loss": 1.1780020904541015,
|
| 351 |
+
"mean_token_accuracy": 0.7016154682636261,
|
| 352 |
+
"num_tokens": 44005399.0,
|
| 353 |
+
"step": 1550
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"entropy": 1.4686012029647828,
|
| 357 |
+
"epoch": 0.54421768707483,
|
| 358 |
+
"grad_norm": 1.0859375,
|
| 359 |
+
"learning_rate": 1.9599074692638007e-05,
|
| 360 |
+
"loss": 1.1757928466796874,
|
| 361 |
+
"mean_token_accuracy": 0.7019401288032532,
|
| 362 |
+
"num_tokens": 45411127.0,
|
| 363 |
+
"step": 1600
|
| 364 |
+
},
|
| 365 |
+
{
|
| 366 |
+
"entropy": 1.4470409440994263,
|
| 367 |
+
"epoch": 0.5612244897959183,
|
| 368 |
+
"grad_norm": 1.234375,
|
| 369 |
+
"learning_rate": 1.9541736425590398e-05,
|
| 370 |
+
"loss": 1.1487213897705078,
|
| 371 |
+
"mean_token_accuracy": 0.708456654548645,
|
| 372 |
+
"num_tokens": 46828771.0,
|
| 373 |
+
"step": 1650
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"entropy": 1.4495110416412353,
|
| 377 |
+
"epoch": 0.5782312925170068,
|
| 378 |
+
"grad_norm": 1.1484375,
|
| 379 |
+
"learning_rate": 1.9480662893720672e-05,
|
| 380 |
+
"loss": 1.1448258972167968,
|
| 381 |
+
"mean_token_accuracy": 0.7081741154193878,
|
| 382 |
+
"num_tokens": 48222827.0,
|
| 383 |
+
"step": 1700
|
| 384 |
+
},
|
| 385 |
+
{
|
| 386 |
+
"entropy": 1.4605646800994874,
|
| 387 |
+
"epoch": 0.5952380952380952,
|
| 388 |
+
"grad_norm": 1.125,
|
| 389 |
+
"learning_rate": 1.9415878005236413e-05,
|
| 390 |
+
"loss": 1.1873452758789063,
|
| 391 |
+
"mean_token_accuracy": 0.6996394228935242,
|
| 392 |
+
"num_tokens": 49670122.0,
|
| 393 |
+
"step": 1750
|
| 394 |
+
},
|
| 395 |
+
{
|
| 396 |
+
"entropy": 1.4478923654556275,
|
| 397 |
+
"epoch": 0.6122448979591837,
|
| 398 |
+
"grad_norm": 1.1484375,
|
| 399 |
+
"learning_rate": 1.9347407121214917e-05,
|
| 400 |
+
"loss": 1.1648704528808593,
|
| 401 |
+
"mean_token_accuracy": 0.7041294479370117,
|
| 402 |
+
"num_tokens": 51095588.0,
|
| 403 |
+
"step": 1800
|
| 404 |
+
},
|
| 405 |
+
{
|
| 406 |
+
"entropy": 1.4468252277374267,
|
| 407 |
+
"epoch": 0.6292517006802721,
|
| 408 |
+
"grad_norm": 1.0703125,
|
| 409 |
+
"learning_rate": 1.927527704567517e-05,
|
| 410 |
+
"loss": 1.1468747711181642,
|
| 411 |
+
"mean_token_accuracy": 0.7075002765655518,
|
| 412 |
+
"num_tokens": 52520201.0,
|
| 413 |
+
"step": 1850
|
| 414 |
+
},
|
| 415 |
+
{
|
| 416 |
+
"entropy": 1.4675183987617493,
|
| 417 |
+
"epoch": 0.6462585034013606,
|
| 418 |
+
"grad_norm": 1.078125,
|
| 419 |
+
"learning_rate": 1.9199516015085025e-05,
|
| 420 |
+
"loss": 1.162491912841797,
|
| 421 |
+
"mean_token_accuracy": 0.7048231554031372,
|
| 422 |
+
"num_tokens": 53944548.0,
|
| 423 |
+
"step": 1900
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"entropy": 1.4320320653915406,
|
| 427 |
+
"epoch": 0.6632653061224489,
|
| 428 |
+
"grad_norm": 1.1640625,
|
| 429 |
+
"learning_rate": 1.912015368730756e-05,
|
| 430 |
+
"loss": 1.1547863006591796,
|
| 431 |
+
"mean_token_accuracy": 0.7064202272891998,
|
| 432 |
+
"num_tokens": 55352104.0,
|
| 433 |
+
"step": 1950
|
| 434 |
+
},
|
| 435 |
+
{
|
| 436 |
+
"entropy": 1.4646633243560792,
|
| 437 |
+
"epoch": 0.6802721088435374,
|
| 438 |
+
"grad_norm": 1.0625,
|
| 439 |
+
"learning_rate": 1.9037221129991045e-05,
|
| 440 |
+
"loss": 1.1570500183105468,
|
| 441 |
+
"mean_token_accuracy": 0.7064848113059997,
|
| 442 |
+
"num_tokens": 56766925.0,
|
| 443 |
+
"step": 2000
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.6802721088435374,
|
| 447 |
+
"eval_entropy": 1.4422022985810994,
|
| 448 |
+
"eval_loss": 1.1238017082214355,
|
| 449 |
+
"eval_mean_token_accuracy": 0.7135469041952566,
|
| 450 |
+
"eval_num_tokens": 56766925.0,
|
| 451 |
+
"eval_runtime": 8.3815,
|
| 452 |
+
"eval_samples_per_second": 113.106,
|
| 453 |
+
"eval_steps_per_second": 14.198,
|
| 454 |
+
"step": 2000
|
| 455 |
+
},
|
| 456 |
+
{
|
| 457 |
+
"entropy": 1.4543561935424805,
|
| 458 |
+
"epoch": 0.6972789115646258,
|
| 459 |
+
"grad_norm": 1.21875,
|
| 460 |
+
"learning_rate": 1.8950750808407016e-05,
|
| 461 |
+
"loss": 1.1518563079833983,
|
| 462 |
+
"mean_token_accuracy": 0.7068261563777923,
|
| 463 |
+
"num_tokens": 58185378.0,
|
| 464 |
+
"step": 2050
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"entropy": 1.4502357959747314,
|
| 468 |
+
"epoch": 0.7142857142857143,
|
| 469 |
+
"grad_norm": 1.0,
|
| 470 |
+
"learning_rate": 1.8860776572741214e-05,
|
| 471 |
+
"loss": 1.1643141174316407,
|
| 472 |
+
"mean_token_accuracy": 0.7048910784721375,
|
| 473 |
+
"num_tokens": 59574639.0,
|
| 474 |
+
"step": 2100
|
| 475 |
+
},
|
| 476 |
+
{
|
| 477 |
+
"entropy": 1.4497376680374146,
|
| 478 |
+
"epoch": 0.7312925170068028,
|
| 479 |
+
"grad_norm": 1.1171875,
|
| 480 |
+
"learning_rate": 1.8767333644842414e-05,
|
| 481 |
+
"loss": 1.1560958862304687,
|
| 482 |
+
"mean_token_accuracy": 0.705675152540207,
|
| 483 |
+
"num_tokens": 60979229.0,
|
| 484 |
+
"step": 2150
|
| 485 |
+
},
|
| 486 |
+
{
|
| 487 |
+
"entropy": 1.4455422043800354,
|
| 488 |
+
"epoch": 0.7482993197278912,
|
| 489 |
+
"grad_norm": 1.1015625,
|
| 490 |
+
"learning_rate": 1.8670458604434253e-05,
|
| 491 |
+
"loss": 1.1813956451416017,
|
| 492 |
+
"mean_token_accuracy": 0.6994379079341888,
|
| 493 |
+
"num_tokens": 62398557.0,
|
| 494 |
+
"step": 2200
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"entropy": 1.4504392910003663,
|
| 498 |
+
"epoch": 0.7653061224489796,
|
| 499 |
+
"grad_norm": 1.2734375,
|
| 500 |
+
"learning_rate": 1.857018937479551e-05,
|
| 501 |
+
"loss": 1.163807601928711,
|
| 502 |
+
"mean_token_accuracy": 0.7050833916664123,
|
| 503 |
+
"num_tokens": 63824693.0,
|
| 504 |
+
"step": 2250
|
| 505 |
+
},
|
| 506 |
+
{
|
| 507 |
+
"entropy": 1.4483773112297058,
|
| 508 |
+
"epoch": 0.782312925170068,
|
| 509 |
+
"grad_norm": 1.03125,
|
| 510 |
+
"learning_rate": 1.846656520791446e-05,
|
| 511 |
+
"loss": 1.1639373779296875,
|
| 512 |
+
"mean_token_accuracy": 0.7042666506767273,
|
| 513 |
+
"num_tokens": 65262483.0,
|
| 514 |
+
"step": 2300
|
| 515 |
+
},
|
| 516 |
+
{
|
| 517 |
+
"entropy": 1.4623826432228089,
|
| 518 |
+
"epoch": 0.7993197278911565,
|
| 519 |
+
"grad_norm": 1.1171875,
|
| 520 |
+
"learning_rate": 1.835962666912303e-05,
|
| 521 |
+
"loss": 1.1614888763427735,
|
| 522 |
+
"mean_token_accuracy": 0.7056748032569885,
|
| 523 |
+
"num_tokens": 66699392.0,
|
| 524 |
+
"step": 2350
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"entropy": 1.4484982657432557,
|
| 528 |
+
"epoch": 0.8163265306122449,
|
| 529 |
+
"grad_norm": 0.94921875,
|
| 530 |
+
"learning_rate": 1.8249415621216853e-05,
|
| 531 |
+
"loss": 1.164166717529297,
|
| 532 |
+
"mean_token_accuracy": 0.7041208696365356,
|
| 533 |
+
"num_tokens": 68163337.0,
|
| 534 |
+
"step": 2400
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"entropy": 1.4403035855293274,
|
| 538 |
+
"epoch": 0.8333333333333334,
|
| 539 |
+
"grad_norm": 1.171875,
|
| 540 |
+
"learning_rate": 1.8135975208067442e-05,
|
| 541 |
+
"loss": 1.1633216857910156,
|
| 542 |
+
"mean_token_accuracy": 0.7039952325820923,
|
| 543 |
+
"num_tokens": 69584260.0,
|
| 544 |
+
"step": 2450
|
| 545 |
+
},
|
| 546 |
+
{
|
| 547 |
+
"entropy": 1.4662260293960572,
|
| 548 |
+
"epoch": 0.8503401360544217,
|
| 549 |
+
"grad_norm": 1.1484375,
|
| 550 |
+
"learning_rate": 1.80193498377328e-05,
|
| 551 |
+
"loss": 1.1636495208740234,
|
| 552 |
+
"mean_token_accuracy": 0.7047837376594543,
|
| 553 |
+
"num_tokens": 70984379.0,
|
| 554 |
+
"step": 2500
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 0.8503401360544217,
|
| 558 |
+
"eval_entropy": 1.4377584517502986,
|
| 559 |
+
"eval_loss": 1.1168923377990723,
|
| 560 |
+
"eval_mean_token_accuracy": 0.7147301846191663,
|
| 561 |
+
"eval_num_tokens": 70984379.0,
|
| 562 |
+
"eval_runtime": 7.3849,
|
| 563 |
+
"eval_samples_per_second": 128.37,
|
| 564 |
+
"eval_steps_per_second": 16.114,
|
| 565 |
+
"step": 2500
|
| 566 |
+
},
|
| 567 |
+
{
|
| 568 |
+
"entropy": 1.443549563884735,
|
| 569 |
+
"epoch": 0.8673469387755102,
|
| 570 |
+
"grad_norm": 1.046875,
|
| 571 |
+
"learning_rate": 1.7899585165073182e-05,
|
| 572 |
+
"loss": 1.148828353881836,
|
| 573 |
+
"mean_token_accuracy": 0.708294941186905,
|
| 574 |
+
"num_tokens": 72388922.0,
|
| 575 |
+
"step": 2550
|
| 576 |
+
},
|
| 577 |
+
{
|
| 578 |
+
"entropy": 1.461836290359497,
|
| 579 |
+
"epoch": 0.8843537414965986,
|
| 580 |
+
"grad_norm": 1.015625,
|
| 581 |
+
"learning_rate": 1.77767280738788e-05,
|
| 582 |
+
"loss": 1.18120361328125,
|
| 583 |
+
"mean_token_accuracy": 0.7006327843666077,
|
| 584 |
+
"num_tokens": 73820763.0,
|
| 585 |
+
"step": 2600
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"entropy": 1.4670860075950622,
|
| 589 |
+
"epoch": 0.9013605442176871,
|
| 590 |
+
"grad_norm": 1.1015625,
|
| 591 |
+
"learning_rate": 1.7650826658516375e-05,
|
| 592 |
+
"loss": 1.190813751220703,
|
| 593 |
+
"mean_token_accuracy": 0.6987618601322174,
|
| 594 |
+
"num_tokens": 75254805.0,
|
| 595 |
+
"step": 2650
|
| 596 |
+
},
|
| 597 |
+
{
|
| 598 |
+
"entropy": 1.4515372467041017,
|
| 599 |
+
"epoch": 0.9183673469387755,
|
| 600 |
+
"grad_norm": 0.9140625,
|
| 601 |
+
"learning_rate": 1.7521930205101857e-05,
|
| 602 |
+
"loss": 1.170417251586914,
|
| 603 |
+
"mean_token_accuracy": 0.7032169091701508,
|
| 604 |
+
"num_tokens": 76693263.0,
|
| 605 |
+
"step": 2700
|
| 606 |
+
},
|
| 607 |
+
{
|
| 608 |
+
"entropy": 1.4385710954666138,
|
| 609 |
+
"epoch": 0.935374149659864,
|
| 610 |
+
"grad_norm": 1.0390625,
|
| 611 |
+
"learning_rate": 1.7390089172206594e-05,
|
| 612 |
+
"loss": 1.1422464752197266,
|
| 613 |
+
"mean_token_accuracy": 0.7097269940376282,
|
| 614 |
+
"num_tokens": 78106463.0,
|
| 615 |
+
"step": 2750
|
| 616 |
+
},
|
| 617 |
+
{
|
| 618 |
+
"entropy": 1.4470832967758178,
|
| 619 |
+
"epoch": 0.9523809523809523,
|
| 620 |
+
"grad_norm": 1.09375,
|
| 621 |
+
"learning_rate": 1.7255355171104492e-05,
|
| 622 |
+
"loss": 1.169587173461914,
|
| 623 |
+
"mean_token_accuracy": 0.7027925884723664,
|
| 624 |
+
"num_tokens": 79532081.0,
|
| 625 |
+
"step": 2800
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"entropy": 1.462675015926361,
|
| 629 |
+
"epoch": 0.9693877551020408,
|
| 630 |
+
"grad_norm": 1.03125,
|
| 631 |
+
"learning_rate": 1.711778094556799e-05,
|
| 632 |
+
"loss": 1.1886711883544923,
|
| 633 |
+
"mean_token_accuracy": 0.6994748151302338,
|
| 634 |
+
"num_tokens": 80999149.0,
|
| 635 |
+
"step": 2850
|
| 636 |
+
},
|
| 637 |
+
{
|
| 638 |
+
"entropy": 1.4251791214942933,
|
| 639 |
+
"epoch": 0.9863945578231292,
|
| 640 |
+
"grad_norm": 0.953125,
|
| 641 |
+
"learning_rate": 1.6977420351220656e-05,
|
| 642 |
+
"loss": 1.138572998046875,
|
| 643 |
+
"mean_token_accuracy": 0.7102291858196259,
|
| 644 |
+
"num_tokens": 82436061.0,
|
| 645 |
+
"step": 2900
|
| 646 |
+
}
|
| 647 |
+
],
|
| 648 |
+
"logging_steps": 50,
|
| 649 |
+
"max_steps": 8820,
|
| 650 |
+
"num_input_tokens_seen": 0,
|
| 651 |
+
"num_train_epochs": 3,
|
| 652 |
+
"save_steps": 500,
|
| 653 |
+
"stateful_callbacks": {
|
| 654 |
+
"TrainerControl": {
|
| 655 |
+
"args": {
|
| 656 |
+
"should_epoch_stop": false,
|
| 657 |
+
"should_evaluate": false,
|
| 658 |
+
"should_log": false,
|
| 659 |
+
"should_save": true,
|
| 660 |
+
"should_training_stop": false
|
| 661 |
+
},
|
| 662 |
+
"attributes": {}
|
| 663 |
+
}
|
| 664 |
+
},
|
| 665 |
+
"total_flos": 4.59538363566528e+17,
|
| 666 |
+
"train_batch_size": 32,
|
| 667 |
+
"trial_name": null,
|
| 668 |
+
"trial_params": null
|
| 669 |
+
}
|
checkpoint-2940/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2309a8eab46e70d77a5e11e1a0025a30065cc758ddf39c81cb9137e50f03b60b
|
| 3 |
+
size 5649
|
checkpoint-5880/config.json
ADDED
|
@@ -0,0 +1,32 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"LlamaForCausalLM"
|
| 4 |
+
],
|
| 5 |
+
"attention_bias": false,
|
| 6 |
+
"attention_dropout": 0.0,
|
| 7 |
+
"bos_token_id": 0,
|
| 8 |
+
"dtype": "bfloat16",
|
| 9 |
+
"eos_token_id": 0,
|
| 10 |
+
"head_dim": 64,
|
| 11 |
+
"hidden_act": "silu",
|
| 12 |
+
"hidden_size": 1280,
|
| 13 |
+
"initializer_range": 0.02,
|
| 14 |
+
"intermediate_size": 2816,
|
| 15 |
+
"max_position_embeddings": 2048,
|
| 16 |
+
"mlp_bias": false,
|
| 17 |
+
"model_type": "llama",
|
| 18 |
+
"num_attention_heads": 20,
|
| 19 |
+
"num_hidden_layers": 24,
|
| 20 |
+
"num_key_value_heads": 10,
|
| 21 |
+
"pad_token_id": 1,
|
| 22 |
+
"pretraining_tp": 1,
|
| 23 |
+
"rms_norm_eps": 1e-05,
|
| 24 |
+
"rope_parameters": {
|
| 25 |
+
"rope_theta": 10000.0,
|
| 26 |
+
"rope_type": "default"
|
| 27 |
+
},
|
| 28 |
+
"tie_word_embeddings": false,
|
| 29 |
+
"transformers_version": "5.3.0",
|
| 30 |
+
"use_cache": false,
|
| 31 |
+
"vocab_size": 50304
|
| 32 |
+
}
|
checkpoint-5880/generation_config.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"bos_token_id": 0,
|
| 4 |
+
"eos_token_id": [
|
| 5 |
+
0
|
| 6 |
+
],
|
| 7 |
+
"pad_token_id": 1,
|
| 8 |
+
"transformers_version": "5.3.0"
|
| 9 |
+
}
|
checkpoint-5880/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2ffacda5bf781914f2fe66f283bee93b041c8459f40108656ea23bdfac1d6050
|
| 3 |
+
size 1012681552
|
checkpoint-5880/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:df5cbd60a73bfa16123116dd3d74a4694d56776e678f0d252559a207cf5dba33
|
| 3 |
+
size 2025501387
|
checkpoint-5880/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:236304ae89e49aae8260113165ee63419b9b745f79120014328a7fa31ed79b42
|
| 3 |
+
size 14645
|
checkpoint-5880/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:94db826fe0bbe58e255f9cd3315d55aa0e79319549b9c17a112721be72d92148
|
| 3 |
+
size 1465
|
checkpoint-5880/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
checkpoint-5880/tokenizer_config.json
ADDED
|
@@ -0,0 +1,21 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_prefix_space": false,
|
| 3 |
+
"backend": "tokenizers",
|
| 4 |
+
"bos_token": "<|endoftext|>",
|
| 5 |
+
"eos_token": "<|endoftext|>",
|
| 6 |
+
"errors": "replace",
|
| 7 |
+
"extra_special_tokens": [
|
| 8 |
+
"<|system|>",
|
| 9 |
+
"<|user|>",
|
| 10 |
+
"<|assistant|>",
|
| 11 |
+
"<|tool_call|>",
|
| 12 |
+
"<|tool_result|>",
|
| 13 |
+
"<|end|>"
|
| 14 |
+
],
|
| 15 |
+
"is_local": true,
|
| 16 |
+
"model_max_length": 1000000000000000019884624838656,
|
| 17 |
+
"pad_token": "<|padding|>",
|
| 18 |
+
"tokenizer_class": "GPTNeoXTokenizer",
|
| 19 |
+
"trim_offsets": true,
|
| 20 |
+
"unk_token": "<|endoftext|>"
|
| 21 |
+
}
|
checkpoint-5880/trainer_state.json
ADDED
|
@@ -0,0 +1,1325 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": null,
|
| 3 |
+
"best_metric": null,
|
| 4 |
+
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 2.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 5880,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"entropy": 1.7698939561843872,
|
| 14 |
+
"epoch": 0.017006802721088437,
|
| 15 |
+
"grad_norm": 3.796875,
|
| 16 |
+
"learning_rate": 1.112372304199773e-06,
|
| 17 |
+
"loss": 1.5452156066894531,
|
| 18 |
+
"mean_token_accuracy": 0.6630511581897736,
|
| 19 |
+
"num_tokens": 1416855.0,
|
| 20 |
+
"step": 50
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"entropy": 1.765206663608551,
|
| 24 |
+
"epoch": 0.034013605442176874,
|
| 25 |
+
"grad_norm": 4.6875,
|
| 26 |
+
"learning_rate": 2.2474460839954596e-06,
|
| 27 |
+
"loss": 1.5233023071289062,
|
| 28 |
+
"mean_token_accuracy": 0.6625268197059632,
|
| 29 |
+
"num_tokens": 2861955.0,
|
| 30 |
+
"step": 100
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"entropy": 1.7640055298805237,
|
| 34 |
+
"epoch": 0.05102040816326531,
|
| 35 |
+
"grad_norm": 2.984375,
|
| 36 |
+
"learning_rate": 3.3825198637911465e-06,
|
| 37 |
+
"loss": 1.4435250854492188,
|
| 38 |
+
"mean_token_accuracy": 0.6652392935752869,
|
| 39 |
+
"num_tokens": 4279472.0,
|
| 40 |
+
"step": 150
|
| 41 |
+
},
|
| 42 |
+
{
|
| 43 |
+
"entropy": 1.7068938755989074,
|
| 44 |
+
"epoch": 0.06802721088435375,
|
| 45 |
+
"grad_norm": 1.3828125,
|
| 46 |
+
"learning_rate": 4.517593643586833e-06,
|
| 47 |
+
"loss": 1.3378631591796875,
|
| 48 |
+
"mean_token_accuracy": 0.673524329662323,
|
| 49 |
+
"num_tokens": 5707775.0,
|
| 50 |
+
"step": 200
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"entropy": 1.6458502221107483,
|
| 54 |
+
"epoch": 0.08503401360544217,
|
| 55 |
+
"grad_norm": 1.1015625,
|
| 56 |
+
"learning_rate": 5.652667423382521e-06,
|
| 57 |
+
"loss": 1.2989051818847657,
|
| 58 |
+
"mean_token_accuracy": 0.6784149742126465,
|
| 59 |
+
"num_tokens": 7118906.0,
|
| 60 |
+
"step": 250
|
| 61 |
+
},
|
| 62 |
+
{
|
| 63 |
+
"entropy": 1.6340991377830505,
|
| 64 |
+
"epoch": 0.10204081632653061,
|
| 65 |
+
"grad_norm": 1.21875,
|
| 66 |
+
"learning_rate": 6.787741203178207e-06,
|
| 67 |
+
"loss": 1.2801470947265625,
|
| 68 |
+
"mean_token_accuracy": 0.6825140416622162,
|
| 69 |
+
"num_tokens": 8529671.0,
|
| 70 |
+
"step": 300
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"entropy": 1.599980125427246,
|
| 74 |
+
"epoch": 0.11904761904761904,
|
| 75 |
+
"grad_norm": 1.1640625,
|
| 76 |
+
"learning_rate": 7.922814982973895e-06,
|
| 77 |
+
"loss": 1.2615231323242186,
|
| 78 |
+
"mean_token_accuracy": 0.6857551598548889,
|
| 79 |
+
"num_tokens": 9973401.0,
|
| 80 |
+
"step": 350
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"entropy": 1.570529818534851,
|
| 84 |
+
"epoch": 0.1360544217687075,
|
| 85 |
+
"grad_norm": 1.078125,
|
| 86 |
+
"learning_rate": 9.057888762769582e-06,
|
| 87 |
+
"loss": 1.2502473449707032,
|
| 88 |
+
"mean_token_accuracy": 0.6881939315795899,
|
| 89 |
+
"num_tokens": 11391611.0,
|
| 90 |
+
"step": 400
|
| 91 |
+
},
|
| 92 |
+
{
|
| 93 |
+
"entropy": 1.5739640021324157,
|
| 94 |
+
"epoch": 0.15306122448979592,
|
| 95 |
+
"grad_norm": 1.1640625,
|
| 96 |
+
"learning_rate": 1.0192962542565269e-05,
|
| 97 |
+
"loss": 1.2381304168701173,
|
| 98 |
+
"mean_token_accuracy": 0.6904172372817993,
|
| 99 |
+
"num_tokens": 12818742.0,
|
| 100 |
+
"step": 450
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"entropy": 1.542729480266571,
|
| 104 |
+
"epoch": 0.17006802721088435,
|
| 105 |
+
"grad_norm": 1.2421875,
|
| 106 |
+
"learning_rate": 1.1328036322360954e-05,
|
| 107 |
+
"loss": 1.2234064483642577,
|
| 108 |
+
"mean_token_accuracy": 0.6928502190113067,
|
| 109 |
+
"num_tokens": 14230694.0,
|
| 110 |
+
"step": 500
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.17006802721088435,
|
| 114 |
+
"eval_entropy": 1.5409097210699771,
|
| 115 |
+
"eval_loss": 1.1968441009521484,
|
| 116 |
+
"eval_mean_token_accuracy": 0.6989319805337602,
|
| 117 |
+
"eval_num_tokens": 14230694.0,
|
| 118 |
+
"eval_runtime": 9.3471,
|
| 119 |
+
"eval_samples_per_second": 101.422,
|
| 120 |
+
"eval_steps_per_second": 12.731,
|
| 121 |
+
"step": 500
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"entropy": 1.5507296323776245,
|
| 125 |
+
"epoch": 0.1870748299319728,
|
| 126 |
+
"grad_norm": 1.2265625,
|
| 127 |
+
"learning_rate": 1.246311010215664e-05,
|
| 128 |
+
"loss": 1.2286612701416015,
|
| 129 |
+
"mean_token_accuracy": 0.6924433386325837,
|
| 130 |
+
"num_tokens": 15656461.0,
|
| 131 |
+
"step": 550
|
| 132 |
+
},
|
| 133 |
+
{
|
| 134 |
+
"entropy": 1.5728167843818666,
|
| 135 |
+
"epoch": 0.20408163265306123,
|
| 136 |
+
"grad_norm": 1.21875,
|
| 137 |
+
"learning_rate": 1.3598183881952328e-05,
|
| 138 |
+
"loss": 1.2478004455566407,
|
| 139 |
+
"mean_token_accuracy": 0.6871815168857575,
|
| 140 |
+
"num_tokens": 17079797.0,
|
| 141 |
+
"step": 600
|
| 142 |
+
},
|
| 143 |
+
{
|
| 144 |
+
"entropy": 1.5617320823669434,
|
| 145 |
+
"epoch": 0.22108843537414966,
|
| 146 |
+
"grad_norm": 1.125,
|
| 147 |
+
"learning_rate": 1.4733257661748014e-05,
|
| 148 |
+
"loss": 1.2422699737548828,
|
| 149 |
+
"mean_token_accuracy": 0.6885153651237488,
|
| 150 |
+
"num_tokens": 18519114.0,
|
| 151 |
+
"step": 650
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"entropy": 1.54343731880188,
|
| 155 |
+
"epoch": 0.23809523809523808,
|
| 156 |
+
"grad_norm": 1.1484375,
|
| 157 |
+
"learning_rate": 1.58683314415437e-05,
|
| 158 |
+
"loss": 1.2279574584960937,
|
| 159 |
+
"mean_token_accuracy": 0.6913824212551117,
|
| 160 |
+
"num_tokens": 19930551.0,
|
| 161 |
+
"step": 700
|
| 162 |
+
},
|
| 163 |
+
{
|
| 164 |
+
"entropy": 1.524279227256775,
|
| 165 |
+
"epoch": 0.25510204081632654,
|
| 166 |
+
"grad_norm": 1.140625,
|
| 167 |
+
"learning_rate": 1.700340522133939e-05,
|
| 168 |
+
"loss": 1.2193751525878906,
|
| 169 |
+
"mean_token_accuracy": 0.6942444491386414,
|
| 170 |
+
"num_tokens": 21345785.0,
|
| 171 |
+
"step": 750
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"entropy": 1.527976930141449,
|
| 175 |
+
"epoch": 0.272108843537415,
|
| 176 |
+
"grad_norm": 1.1171875,
|
| 177 |
+
"learning_rate": 1.8138479001135073e-05,
|
| 178 |
+
"loss": 1.212171859741211,
|
| 179 |
+
"mean_token_accuracy": 0.6960112357139587,
|
| 180 |
+
"num_tokens": 22754238.0,
|
| 181 |
+
"step": 800
|
| 182 |
+
},
|
| 183 |
+
{
|
| 184 |
+
"entropy": 1.5164598178863526,
|
| 185 |
+
"epoch": 0.2891156462585034,
|
| 186 |
+
"grad_norm": 1.109375,
|
| 187 |
+
"learning_rate": 1.9273552780930764e-05,
|
| 188 |
+
"loss": 1.2044546508789062,
|
| 189 |
+
"mean_token_accuracy": 0.6967348265647888,
|
| 190 |
+
"num_tokens": 24167833.0,
|
| 191 |
+
"step": 850
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"entropy": 1.4968056178092957,
|
| 195 |
+
"epoch": 0.30612244897959184,
|
| 196 |
+
"grad_norm": 1.0234375,
|
| 197 |
+
"learning_rate": 1.9999746322867523e-05,
|
| 198 |
+
"loss": 1.2030063629150392,
|
| 199 |
+
"mean_token_accuracy": 0.6967223298549652,
|
| 200 |
+
"num_tokens": 25597839.0,
|
| 201 |
+
"step": 900
|
| 202 |
+
},
|
| 203 |
+
{
|
| 204 |
+
"entropy": 1.5187969017028808,
|
| 205 |
+
"epoch": 0.3231292517006803,
|
| 206 |
+
"grad_norm": 1.1796875,
|
| 207 |
+
"learning_rate": 1.9996379823325586e-05,
|
| 208 |
+
"loss": 1.2039017486572265,
|
| 209 |
+
"mean_token_accuracy": 0.69672048330307,
|
| 210 |
+
"num_tokens": 27039491.0,
|
| 211 |
+
"step": 950
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"entropy": 1.5000163841247558,
|
| 215 |
+
"epoch": 0.3401360544217687,
|
| 216 |
+
"grad_norm": 1.0546875,
|
| 217 |
+
"learning_rate": 1.9989100081556328e-05,
|
| 218 |
+
"loss": 1.2096540832519531,
|
| 219 |
+
"mean_token_accuracy": 0.696036319732666,
|
| 220 |
+
"num_tokens": 28462525.0,
|
| 221 |
+
"step": 1000
|
| 222 |
+
},
|
| 223 |
+
{
|
| 224 |
+
"epoch": 0.3401360544217687,
|
| 225 |
+
"eval_entropy": 1.4825329389892707,
|
| 226 |
+
"eval_loss": 1.1529995203018188,
|
| 227 |
+
"eval_mean_token_accuracy": 0.7072621358542883,
|
| 228 |
+
"eval_num_tokens": 28462525.0,
|
| 229 |
+
"eval_runtime": 9.42,
|
| 230 |
+
"eval_samples_per_second": 100.637,
|
| 231 |
+
"eval_steps_per_second": 12.633,
|
| 232 |
+
"step": 1000
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"entropy": 1.4857492208480836,
|
| 236 |
+
"epoch": 0.35714285714285715,
|
| 237 |
+
"grad_norm": 1.0625,
|
| 238 |
+
"learning_rate": 1.997790994733071e-05,
|
| 239 |
+
"loss": 1.1972512054443358,
|
| 240 |
+
"mean_token_accuracy": 0.6976107740402222,
|
| 241 |
+
"num_tokens": 29880725.0,
|
| 242 |
+
"step": 1050
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"entropy": 1.4679273128509522,
|
| 246 |
+
"epoch": 0.3741496598639456,
|
| 247 |
+
"grad_norm": 1.125,
|
| 248 |
+
"learning_rate": 1.9962813801205148e-05,
|
| 249 |
+
"loss": 1.1646170806884766,
|
| 250 |
+
"mean_token_accuracy": 0.705207599401474,
|
| 251 |
+
"num_tokens": 31288250.0,
|
| 252 |
+
"step": 1100
|
| 253 |
+
},
|
| 254 |
+
{
|
| 255 |
+
"entropy": 1.482924530506134,
|
| 256 |
+
"epoch": 0.391156462585034,
|
| 257 |
+
"grad_norm": 1.25,
|
| 258 |
+
"learning_rate": 1.9943817552806678e-05,
|
| 259 |
+
"loss": 1.1940898132324218,
|
| 260 |
+
"mean_token_accuracy": 0.6988638079166413,
|
| 261 |
+
"num_tokens": 32696424.0,
|
| 262 |
+
"step": 1150
|
| 263 |
+
},
|
| 264 |
+
{
|
| 265 |
+
"entropy": 1.4768511247634888,
|
| 266 |
+
"epoch": 0.40816326530612246,
|
| 267 |
+
"grad_norm": 1.1015625,
|
| 268 |
+
"learning_rate": 1.9920928638519545e-05,
|
| 269 |
+
"loss": 1.1782675933837892,
|
| 270 |
+
"mean_token_accuracy": 0.7024707806110382,
|
| 271 |
+
"num_tokens": 34115752.0,
|
| 272 |
+
"step": 1200
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"entropy": 1.4849171042442322,
|
| 276 |
+
"epoch": 0.42517006802721086,
|
| 277 |
+
"grad_norm": 1.1015625,
|
| 278 |
+
"learning_rate": 1.9894156018574104e-05,
|
| 279 |
+
"loss": 1.1730673217773437,
|
| 280 |
+
"mean_token_accuracy": 0.7031274008750915,
|
| 281 |
+
"num_tokens": 35496076.0,
|
| 282 |
+
"step": 1250
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"entropy": 1.477535102367401,
|
| 286 |
+
"epoch": 0.4421768707482993,
|
| 287 |
+
"grad_norm": 1.140625,
|
| 288 |
+
"learning_rate": 1.9863510173539193e-05,
|
| 289 |
+
"loss": 1.1849075317382813,
|
| 290 |
+
"mean_token_accuracy": 0.6999362814426422,
|
| 291 |
+
"num_tokens": 36924227.0,
|
| 292 |
+
"step": 1300
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"entropy": 1.4699360299110413,
|
| 296 |
+
"epoch": 0.45918367346938777,
|
| 297 |
+
"grad_norm": 0.9921875,
|
| 298 |
+
"learning_rate": 1.982900310021936e-05,
|
| 299 |
+
"loss": 1.1758278656005858,
|
| 300 |
+
"mean_token_accuracy": 0.7016613483428955,
|
| 301 |
+
"num_tokens": 38298695.0,
|
| 302 |
+
"step": 1350
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"entropy": 1.4796976780891418,
|
| 306 |
+
"epoch": 0.47619047619047616,
|
| 307 |
+
"grad_norm": 1.2421875,
|
| 308 |
+
"learning_rate": 1.9790648306958504e-05,
|
| 309 |
+
"loss": 1.1764393615722657,
|
| 310 |
+
"mean_token_accuracy": 0.7006231868267059,
|
| 311 |
+
"num_tokens": 39730904.0,
|
| 312 |
+
"step": 1400
|
| 313 |
+
},
|
| 314 |
+
{
|
| 315 |
+
"entropy": 1.4659403300285339,
|
| 316 |
+
"epoch": 0.4931972789115646,
|
| 317 |
+
"grad_norm": 1.109375,
|
| 318 |
+
"learning_rate": 1.974846080835184e-05,
|
| 319 |
+
"loss": 1.1739279174804687,
|
| 320 |
+
"mean_token_accuracy": 0.7042246878147125,
|
| 321 |
+
"num_tokens": 41159605.0,
|
| 322 |
+
"step": 1450
|
| 323 |
+
},
|
| 324 |
+
{
|
| 325 |
+
"entropy": 1.4660640573501587,
|
| 326 |
+
"epoch": 0.5102040816326531,
|
| 327 |
+
"grad_norm": 1.15625,
|
| 328 |
+
"learning_rate": 1.9702457119368175e-05,
|
| 329 |
+
"loss": 1.1742501831054688,
|
| 330 |
+
"mean_token_accuracy": 0.7028728783130646,
|
| 331 |
+
"num_tokens": 42591337.0,
|
| 332 |
+
"step": 1500
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 0.5102040816326531,
|
| 336 |
+
"eval_entropy": 1.4500642824573677,
|
| 337 |
+
"eval_loss": 1.1341153383255005,
|
| 338 |
+
"eval_mean_token_accuracy": 0.7110618072397569,
|
| 339 |
+
"eval_num_tokens": 42591337.0,
|
| 340 |
+
"eval_runtime": 9.205,
|
| 341 |
+
"eval_samples_per_second": 102.988,
|
| 342 |
+
"eval_steps_per_second": 12.928,
|
| 343 |
+
"step": 1500
|
| 344 |
+
},
|
| 345 |
+
{
|
| 346 |
+
"entropy": 1.4817289400100708,
|
| 347 |
+
"epoch": 0.5272108843537415,
|
| 348 |
+
"grad_norm": 1.15625,
|
| 349 |
+
"learning_rate": 1.965265524888488e-05,
|
| 350 |
+
"loss": 1.1780020904541015,
|
| 351 |
+
"mean_token_accuracy": 0.7016154682636261,
|
| 352 |
+
"num_tokens": 44005399.0,
|
| 353 |
+
"step": 1550
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"entropy": 1.4686012029647828,
|
| 357 |
+
"epoch": 0.54421768707483,
|
| 358 |
+
"grad_norm": 1.0859375,
|
| 359 |
+
"learning_rate": 1.9599074692638007e-05,
|
| 360 |
+
"loss": 1.1757928466796874,
|
| 361 |
+
"mean_token_accuracy": 0.7019401288032532,
|
| 362 |
+
"num_tokens": 45411127.0,
|
| 363 |
+
"step": 1600
|
| 364 |
+
},
|
| 365 |
+
{
|
| 366 |
+
"entropy": 1.4470409440994263,
|
| 367 |
+
"epoch": 0.5612244897959183,
|
| 368 |
+
"grad_norm": 1.234375,
|
| 369 |
+
"learning_rate": 1.9541736425590398e-05,
|
| 370 |
+
"loss": 1.1487213897705078,
|
| 371 |
+
"mean_token_accuracy": 0.708456654548645,
|
| 372 |
+
"num_tokens": 46828771.0,
|
| 373 |
+
"step": 1650
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"entropy": 1.4495110416412353,
|
| 377 |
+
"epoch": 0.5782312925170068,
|
| 378 |
+
"grad_norm": 1.1484375,
|
| 379 |
+
"learning_rate": 1.9480662893720672e-05,
|
| 380 |
+
"loss": 1.1448258972167968,
|
| 381 |
+
"mean_token_accuracy": 0.7081741154193878,
|
| 382 |
+
"num_tokens": 48222827.0,
|
| 383 |
+
"step": 1700
|
| 384 |
+
},
|
| 385 |
+
{
|
| 386 |
+
"entropy": 1.4605646800994874,
|
| 387 |
+
"epoch": 0.5952380952380952,
|
| 388 |
+
"grad_norm": 1.125,
|
| 389 |
+
"learning_rate": 1.9415878005236413e-05,
|
| 390 |
+
"loss": 1.1873452758789063,
|
| 391 |
+
"mean_token_accuracy": 0.6996394228935242,
|
| 392 |
+
"num_tokens": 49670122.0,
|
| 393 |
+
"step": 1750
|
| 394 |
+
},
|
| 395 |
+
{
|
| 396 |
+
"entropy": 1.4478923654556275,
|
| 397 |
+
"epoch": 0.6122448979591837,
|
| 398 |
+
"grad_norm": 1.1484375,
|
| 399 |
+
"learning_rate": 1.9347407121214917e-05,
|
| 400 |
+
"loss": 1.1648704528808593,
|
| 401 |
+
"mean_token_accuracy": 0.7041294479370117,
|
| 402 |
+
"num_tokens": 51095588.0,
|
| 403 |
+
"step": 1800
|
| 404 |
+
},
|
| 405 |
+
{
|
| 406 |
+
"entropy": 1.4468252277374267,
|
| 407 |
+
"epoch": 0.6292517006802721,
|
| 408 |
+
"grad_norm": 1.0703125,
|
| 409 |
+
"learning_rate": 1.927527704567517e-05,
|
| 410 |
+
"loss": 1.1468747711181642,
|
| 411 |
+
"mean_token_accuracy": 0.7075002765655518,
|
| 412 |
+
"num_tokens": 52520201.0,
|
| 413 |
+
"step": 1850
|
| 414 |
+
},
|
| 415 |
+
{
|
| 416 |
+
"entropy": 1.4675183987617493,
|
| 417 |
+
"epoch": 0.6462585034013606,
|
| 418 |
+
"grad_norm": 1.078125,
|
| 419 |
+
"learning_rate": 1.9199516015085025e-05,
|
| 420 |
+
"loss": 1.162491912841797,
|
| 421 |
+
"mean_token_accuracy": 0.7048231554031372,
|
| 422 |
+
"num_tokens": 53944548.0,
|
| 423 |
+
"step": 1900
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"entropy": 1.4320320653915406,
|
| 427 |
+
"epoch": 0.6632653061224489,
|
| 428 |
+
"grad_norm": 1.1640625,
|
| 429 |
+
"learning_rate": 1.912015368730756e-05,
|
| 430 |
+
"loss": 1.1547863006591796,
|
| 431 |
+
"mean_token_accuracy": 0.7064202272891998,
|
| 432 |
+
"num_tokens": 55352104.0,
|
| 433 |
+
"step": 1950
|
| 434 |
+
},
|
| 435 |
+
{
|
| 436 |
+
"entropy": 1.4646633243560792,
|
| 437 |
+
"epoch": 0.6802721088435374,
|
| 438 |
+
"grad_norm": 1.0625,
|
| 439 |
+
"learning_rate": 1.9037221129991045e-05,
|
| 440 |
+
"loss": 1.1570500183105468,
|
| 441 |
+
"mean_token_accuracy": 0.7064848113059997,
|
| 442 |
+
"num_tokens": 56766925.0,
|
| 443 |
+
"step": 2000
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.6802721088435374,
|
| 447 |
+
"eval_entropy": 1.4422022985810994,
|
| 448 |
+
"eval_loss": 1.1238017082214355,
|
| 449 |
+
"eval_mean_token_accuracy": 0.7135469041952566,
|
| 450 |
+
"eval_num_tokens": 56766925.0,
|
| 451 |
+
"eval_runtime": 8.3815,
|
| 452 |
+
"eval_samples_per_second": 113.106,
|
| 453 |
+
"eval_steps_per_second": 14.198,
|
| 454 |
+
"step": 2000
|
| 455 |
+
},
|
| 456 |
+
{
|
| 457 |
+
"entropy": 1.4543561935424805,
|
| 458 |
+
"epoch": 0.6972789115646258,
|
| 459 |
+
"grad_norm": 1.21875,
|
| 460 |
+
"learning_rate": 1.8950750808407016e-05,
|
| 461 |
+
"loss": 1.1518563079833983,
|
| 462 |
+
"mean_token_accuracy": 0.7068261563777923,
|
| 463 |
+
"num_tokens": 58185378.0,
|
| 464 |
+
"step": 2050
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"entropy": 1.4502357959747314,
|
| 468 |
+
"epoch": 0.7142857142857143,
|
| 469 |
+
"grad_norm": 1.0,
|
| 470 |
+
"learning_rate": 1.8860776572741214e-05,
|
| 471 |
+
"loss": 1.1643141174316407,
|
| 472 |
+
"mean_token_accuracy": 0.7048910784721375,
|
| 473 |
+
"num_tokens": 59574639.0,
|
| 474 |
+
"step": 2100
|
| 475 |
+
},
|
| 476 |
+
{
|
| 477 |
+
"entropy": 1.4497376680374146,
|
| 478 |
+
"epoch": 0.7312925170068028,
|
| 479 |
+
"grad_norm": 1.1171875,
|
| 480 |
+
"learning_rate": 1.8767333644842414e-05,
|
| 481 |
+
"loss": 1.1560958862304687,
|
| 482 |
+
"mean_token_accuracy": 0.705675152540207,
|
| 483 |
+
"num_tokens": 60979229.0,
|
| 484 |
+
"step": 2150
|
| 485 |
+
},
|
| 486 |
+
{
|
| 487 |
+
"entropy": 1.4455422043800354,
|
| 488 |
+
"epoch": 0.7482993197278912,
|
| 489 |
+
"grad_norm": 1.1015625,
|
| 490 |
+
"learning_rate": 1.8670458604434253e-05,
|
| 491 |
+
"loss": 1.1813956451416017,
|
| 492 |
+
"mean_token_accuracy": 0.6994379079341888,
|
| 493 |
+
"num_tokens": 62398557.0,
|
| 494 |
+
"step": 2200
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"entropy": 1.4504392910003663,
|
| 498 |
+
"epoch": 0.7653061224489796,
|
| 499 |
+
"grad_norm": 1.2734375,
|
| 500 |
+
"learning_rate": 1.857018937479551e-05,
|
| 501 |
+
"loss": 1.163807601928711,
|
| 502 |
+
"mean_token_accuracy": 0.7050833916664123,
|
| 503 |
+
"num_tokens": 63824693.0,
|
| 504 |
+
"step": 2250
|
| 505 |
+
},
|
| 506 |
+
{
|
| 507 |
+
"entropy": 1.4483773112297058,
|
| 508 |
+
"epoch": 0.782312925170068,
|
| 509 |
+
"grad_norm": 1.03125,
|
| 510 |
+
"learning_rate": 1.846656520791446e-05,
|
| 511 |
+
"loss": 1.1639373779296875,
|
| 512 |
+
"mean_token_accuracy": 0.7042666506767273,
|
| 513 |
+
"num_tokens": 65262483.0,
|
| 514 |
+
"step": 2300
|
| 515 |
+
},
|
| 516 |
+
{
|
| 517 |
+
"entropy": 1.4623826432228089,
|
| 518 |
+
"epoch": 0.7993197278911565,
|
| 519 |
+
"grad_norm": 1.1171875,
|
| 520 |
+
"learning_rate": 1.835962666912303e-05,
|
| 521 |
+
"loss": 1.1614888763427735,
|
| 522 |
+
"mean_token_accuracy": 0.7056748032569885,
|
| 523 |
+
"num_tokens": 66699392.0,
|
| 524 |
+
"step": 2350
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"entropy": 1.4484982657432557,
|
| 528 |
+
"epoch": 0.8163265306122449,
|
| 529 |
+
"grad_norm": 0.94921875,
|
| 530 |
+
"learning_rate": 1.8249415621216853e-05,
|
| 531 |
+
"loss": 1.164166717529297,
|
| 532 |
+
"mean_token_accuracy": 0.7041208696365356,
|
| 533 |
+
"num_tokens": 68163337.0,
|
| 534 |
+
"step": 2400
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"entropy": 1.4403035855293274,
|
| 538 |
+
"epoch": 0.8333333333333334,
|
| 539 |
+
"grad_norm": 1.171875,
|
| 540 |
+
"learning_rate": 1.8135975208067442e-05,
|
| 541 |
+
"loss": 1.1633216857910156,
|
| 542 |
+
"mean_token_accuracy": 0.7039952325820923,
|
| 543 |
+
"num_tokens": 69584260.0,
|
| 544 |
+
"step": 2450
|
| 545 |
+
},
|
| 546 |
+
{
|
| 547 |
+
"entropy": 1.4662260293960572,
|
| 548 |
+
"epoch": 0.8503401360544217,
|
| 549 |
+
"grad_norm": 1.1484375,
|
| 550 |
+
"learning_rate": 1.80193498377328e-05,
|
| 551 |
+
"loss": 1.1636495208740234,
|
| 552 |
+
"mean_token_accuracy": 0.7047837376594543,
|
| 553 |
+
"num_tokens": 70984379.0,
|
| 554 |
+
"step": 2500
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 0.8503401360544217,
|
| 558 |
+
"eval_entropy": 1.4377584517502986,
|
| 559 |
+
"eval_loss": 1.1168923377990723,
|
| 560 |
+
"eval_mean_token_accuracy": 0.7147301846191663,
|
| 561 |
+
"eval_num_tokens": 70984379.0,
|
| 562 |
+
"eval_runtime": 7.3849,
|
| 563 |
+
"eval_samples_per_second": 128.37,
|
| 564 |
+
"eval_steps_per_second": 16.114,
|
| 565 |
+
"step": 2500
|
| 566 |
+
},
|
| 567 |
+
{
|
| 568 |
+
"entropy": 1.443549563884735,
|
| 569 |
+
"epoch": 0.8673469387755102,
|
| 570 |
+
"grad_norm": 1.046875,
|
| 571 |
+
"learning_rate": 1.7899585165073182e-05,
|
| 572 |
+
"loss": 1.148828353881836,
|
| 573 |
+
"mean_token_accuracy": 0.708294941186905,
|
| 574 |
+
"num_tokens": 72388922.0,
|
| 575 |
+
"step": 2550
|
| 576 |
+
},
|
| 577 |
+
{
|
| 578 |
+
"entropy": 1.461836290359497,
|
| 579 |
+
"epoch": 0.8843537414965986,
|
| 580 |
+
"grad_norm": 1.015625,
|
| 581 |
+
"learning_rate": 1.77767280738788e-05,
|
| 582 |
+
"loss": 1.18120361328125,
|
| 583 |
+
"mean_token_accuracy": 0.7006327843666077,
|
| 584 |
+
"num_tokens": 73820763.0,
|
| 585 |
+
"step": 2600
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"entropy": 1.4670860075950622,
|
| 589 |
+
"epoch": 0.9013605442176871,
|
| 590 |
+
"grad_norm": 1.1015625,
|
| 591 |
+
"learning_rate": 1.7650826658516375e-05,
|
| 592 |
+
"loss": 1.190813751220703,
|
| 593 |
+
"mean_token_accuracy": 0.6987618601322174,
|
| 594 |
+
"num_tokens": 75254805.0,
|
| 595 |
+
"step": 2650
|
| 596 |
+
},
|
| 597 |
+
{
|
| 598 |
+
"entropy": 1.4515372467041017,
|
| 599 |
+
"epoch": 0.9183673469387755,
|
| 600 |
+
"grad_norm": 0.9140625,
|
| 601 |
+
"learning_rate": 1.7521930205101857e-05,
|
| 602 |
+
"loss": 1.170417251586914,
|
| 603 |
+
"mean_token_accuracy": 0.7032169091701508,
|
| 604 |
+
"num_tokens": 76693263.0,
|
| 605 |
+
"step": 2700
|
| 606 |
+
},
|
| 607 |
+
{
|
| 608 |
+
"entropy": 1.4385710954666138,
|
| 609 |
+
"epoch": 0.935374149659864,
|
| 610 |
+
"grad_norm": 1.0390625,
|
| 611 |
+
"learning_rate": 1.7390089172206594e-05,
|
| 612 |
+
"loss": 1.1422464752197266,
|
| 613 |
+
"mean_token_accuracy": 0.7097269940376282,
|
| 614 |
+
"num_tokens": 78106463.0,
|
| 615 |
+
"step": 2750
|
| 616 |
+
},
|
| 617 |
+
{
|
| 618 |
+
"entropy": 1.4470832967758178,
|
| 619 |
+
"epoch": 0.9523809523809523,
|
| 620 |
+
"grad_norm": 1.09375,
|
| 621 |
+
"learning_rate": 1.7255355171104492e-05,
|
| 622 |
+
"loss": 1.169587173461914,
|
| 623 |
+
"mean_token_accuracy": 0.7027925884723664,
|
| 624 |
+
"num_tokens": 79532081.0,
|
| 625 |
+
"step": 2800
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"entropy": 1.462675015926361,
|
| 629 |
+
"epoch": 0.9693877551020408,
|
| 630 |
+
"grad_norm": 1.03125,
|
| 631 |
+
"learning_rate": 1.711778094556799e-05,
|
| 632 |
+
"loss": 1.1886711883544923,
|
| 633 |
+
"mean_token_accuracy": 0.6994748151302338,
|
| 634 |
+
"num_tokens": 80999149.0,
|
| 635 |
+
"step": 2850
|
| 636 |
+
},
|
| 637 |
+
{
|
| 638 |
+
"entropy": 1.4251791214942933,
|
| 639 |
+
"epoch": 0.9863945578231292,
|
| 640 |
+
"grad_norm": 0.953125,
|
| 641 |
+
"learning_rate": 1.6977420351220656e-05,
|
| 642 |
+
"loss": 1.138572998046875,
|
| 643 |
+
"mean_token_accuracy": 0.7102291858196259,
|
| 644 |
+
"num_tokens": 82436061.0,
|
| 645 |
+
"step": 2900
|
| 646 |
+
},
|
| 647 |
+
{
|
| 648 |
+
"entropy": 1.4440323853492736,
|
| 649 |
+
"epoch": 1.0034013605442176,
|
| 650 |
+
"grad_norm": 1.171875,
|
| 651 |
+
"learning_rate": 1.683432833445453e-05,
|
| 652 |
+
"loss": 1.1582012176513672,
|
| 653 |
+
"mean_token_accuracy": 0.7026605701446533,
|
| 654 |
+
"num_tokens": 83831706.0,
|
| 655 |
+
"step": 2950
|
| 656 |
+
},
|
| 657 |
+
{
|
| 658 |
+
"entropy": 1.4328801035881042,
|
| 659 |
+
"epoch": 1.0204081632653061,
|
| 660 |
+
"grad_norm": 1.046875,
|
| 661 |
+
"learning_rate": 1.668856091092051e-05,
|
| 662 |
+
"loss": 1.1449398040771483,
|
| 663 |
+
"mean_token_accuracy": 0.7084557473659515,
|
| 664 |
+
"num_tokens": 85243788.0,
|
| 665 |
+
"step": 3000
|
| 666 |
+
},
|
| 667 |
+
{
|
| 668 |
+
"epoch": 1.0204081632653061,
|
| 669 |
+
"eval_entropy": 1.425075779442026,
|
| 670 |
+
"eval_loss": 1.112339973449707,
|
| 671 |
+
"eval_mean_token_accuracy": 0.7156877617876068,
|
| 672 |
+
"eval_num_tokens": 85243788.0,
|
| 673 |
+
"eval_runtime": 9.148,
|
| 674 |
+
"eval_samples_per_second": 103.629,
|
| 675 |
+
"eval_steps_per_second": 13.008,
|
| 676 |
+
"step": 3000
|
| 677 |
+
},
|
| 678 |
+
{
|
| 679 |
+
"entropy": 1.457193055152893,
|
| 680 |
+
"epoch": 1.0374149659863945,
|
| 681 |
+
"grad_norm": 1.0078125,
|
| 682 |
+
"learning_rate": 1.6540175143600124e-05,
|
| 683 |
+
"loss": 1.1573681640625,
|
| 684 |
+
"mean_token_accuracy": 0.706193675994873,
|
| 685 |
+
"num_tokens": 86631509.0,
|
| 686 |
+
"step": 3050
|
| 687 |
+
},
|
| 688 |
+
{
|
| 689 |
+
"entropy": 1.4502310729026795,
|
| 690 |
+
"epoch": 1.054421768707483,
|
| 691 |
+
"grad_norm": 1.125,
|
| 692 |
+
"learning_rate": 1.6389229120467304e-05,
|
| 693 |
+
"loss": 1.1431642913818358,
|
| 694 |
+
"mean_token_accuracy": 0.7098660218715668,
|
| 695 |
+
"num_tokens": 88045606.0,
|
| 696 |
+
"step": 3100
|
| 697 |
+
},
|
| 698 |
+
{
|
| 699 |
+
"entropy": 1.4366277551651,
|
| 700 |
+
"epoch": 1.0714285714285714,
|
| 701 |
+
"grad_norm": 0.9453125,
|
| 702 |
+
"learning_rate": 1.6235781931748923e-05,
|
| 703 |
+
"loss": 1.1489845275878907,
|
| 704 |
+
"mean_token_accuracy": 0.707208307981491,
|
| 705 |
+
"num_tokens": 89467532.0,
|
| 706 |
+
"step": 3150
|
| 707 |
+
},
|
| 708 |
+
{
|
| 709 |
+
"entropy": 1.4366546082496643,
|
| 710 |
+
"epoch": 1.08843537414966,
|
| 711 |
+
"grad_norm": 1.1875,
|
| 712 |
+
"learning_rate": 1.6079893646793016e-05,
|
| 713 |
+
"loss": 1.1411451721191406,
|
| 714 |
+
"mean_token_accuracy": 0.7094867730140686,
|
| 715 |
+
"num_tokens": 90884635.0,
|
| 716 |
+
"step": 3200
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"entropy": 1.451173803806305,
|
| 720 |
+
"epoch": 1.1054421768707483,
|
| 721 |
+
"grad_norm": 0.98828125,
|
| 722 |
+
"learning_rate": 1.5921625290553628e-05,
|
| 723 |
+
"loss": 1.168065185546875,
|
| 724 |
+
"mean_token_accuracy": 0.7023441231250763,
|
| 725 |
+
"num_tokens": 92294540.0,
|
| 726 |
+
"step": 3250
|
| 727 |
+
},
|
| 728 |
+
{
|
| 729 |
+
"entropy": 1.4395431089401245,
|
| 730 |
+
"epoch": 1.1224489795918366,
|
| 731 |
+
"grad_norm": 1.0703125,
|
| 732 |
+
"learning_rate": 1.5761038819701658e-05,
|
| 733 |
+
"loss": 1.1530582427978515,
|
| 734 |
+
"mean_token_accuracy": 0.7062457311153412,
|
| 735 |
+
"num_tokens": 93722473.0,
|
| 736 |
+
"step": 3300
|
| 737 |
+
},
|
| 738 |
+
{
|
| 739 |
+
"entropy": 1.4319015264511108,
|
| 740 |
+
"epoch": 1.1394557823129252,
|
| 741 |
+
"grad_norm": 1.078125,
|
| 742 |
+
"learning_rate": 1.5598197098370928e-05,
|
| 743 |
+
"loss": 1.135885009765625,
|
| 744 |
+
"mean_token_accuracy": 0.7098275601863862,
|
| 745 |
+
"num_tokens": 95130619.0,
|
| 746 |
+
"step": 3350
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"entropy": 1.4478510427474975,
|
| 750 |
+
"epoch": 1.1564625850340136,
|
| 751 |
+
"grad_norm": 1.125,
|
| 752 |
+
"learning_rate": 1.5433163873548994e-05,
|
| 753 |
+
"loss": 1.1470938110351563,
|
| 754 |
+
"mean_token_accuracy": 0.7073779857158661,
|
| 755 |
+
"num_tokens": 96554760.0,
|
| 756 |
+
"step": 3400
|
| 757 |
+
},
|
| 758 |
+
{
|
| 759 |
+
"entropy": 1.4374464821815491,
|
| 760 |
+
"epoch": 1.1734693877551021,
|
| 761 |
+
"grad_norm": 1.1328125,
|
| 762 |
+
"learning_rate": 1.5266003750122394e-05,
|
| 763 |
+
"loss": 1.1405145263671874,
|
| 764 |
+
"mean_token_accuracy": 0.7093812191486358,
|
| 765 |
+
"num_tokens": 97955334.0,
|
| 766 |
+
"step": 3450
|
| 767 |
+
},
|
| 768 |
+
{
|
| 769 |
+
"entropy": 1.4236741828918458,
|
| 770 |
+
"epoch": 1.1904761904761905,
|
| 771 |
+
"grad_norm": 1.046875,
|
| 772 |
+
"learning_rate": 1.5096782165586037e-05,
|
| 773 |
+
"loss": 1.1330892181396484,
|
| 774 |
+
"mean_token_accuracy": 0.7110612964630127,
|
| 775 |
+
"num_tokens": 99395276.0,
|
| 776 |
+
"step": 3500
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 1.1904761904761905,
|
| 780 |
+
"eval_entropy": 1.4220578530255485,
|
| 781 |
+
"eval_loss": 1.1096009016036987,
|
| 782 |
+
"eval_mean_token_accuracy": 0.7162127835409982,
|
| 783 |
+
"eval_num_tokens": 99395276.0,
|
| 784 |
+
"eval_runtime": 7.5607,
|
| 785 |
+
"eval_samples_per_second": 125.385,
|
| 786 |
+
"eval_steps_per_second": 15.739,
|
| 787 |
+
"step": 3500
|
| 788 |
+
},
|
| 789 |
+
{
|
| 790 |
+
"entropy": 1.4502542996406556,
|
| 791 |
+
"epoch": 1.2074829931972788,
|
| 792 |
+
"grad_norm": 1.109375,
|
| 793 |
+
"learning_rate": 1.4925565364426626e-05,
|
| 794 |
+
"loss": 1.1476077270507812,
|
| 795 |
+
"mean_token_accuracy": 0.7080324268341065,
|
| 796 |
+
"num_tokens": 100822779.0,
|
| 797 |
+
"step": 3550
|
| 798 |
+
},
|
| 799 |
+
{
|
| 800 |
+
"entropy": 1.4335950636863708,
|
| 801 |
+
"epoch": 1.2244897959183674,
|
| 802 |
+
"grad_norm": 1.015625,
|
| 803 |
+
"learning_rate": 1.4752420372190235e-05,
|
| 804 |
+
"loss": 1.1536802673339843,
|
| 805 |
+
"mean_token_accuracy": 0.7067959046363831,
|
| 806 |
+
"num_tokens": 102255057.0,
|
| 807 |
+
"step": 3600
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"entropy": 1.431325912475586,
|
| 811 |
+
"epoch": 1.2414965986394557,
|
| 812 |
+
"grad_norm": 1.1171875,
|
| 813 |
+
"learning_rate": 1.4577414969244052e-05,
|
| 814 |
+
"loss": 1.1357704162597657,
|
| 815 |
+
"mean_token_accuracy": 0.7090205514431,
|
| 816 |
+
"num_tokens": 103682059.0,
|
| 817 |
+
"step": 3650
|
| 818 |
+
},
|
| 819 |
+
{
|
| 820 |
+
"entropy": 1.430570056438446,
|
| 821 |
+
"epoch": 1.2585034013605443,
|
| 822 |
+
"grad_norm": 1.0625,
|
| 823 |
+
"learning_rate": 1.4400617664242706e-05,
|
| 824 |
+
"loss": 1.1330258178710937,
|
| 825 |
+
"mean_token_accuracy": 0.7101845383644104,
|
| 826 |
+
"num_tokens": 105094949.0,
|
| 827 |
+
"step": 3700
|
| 828 |
+
},
|
| 829 |
+
{
|
| 830 |
+
"entropy": 1.4420754599571228,
|
| 831 |
+
"epoch": 1.2755102040816326,
|
| 832 |
+
"grad_norm": 0.96875,
|
| 833 |
+
"learning_rate": 1.422209766730945e-05,
|
| 834 |
+
"loss": 1.1578455352783203,
|
| 835 |
+
"mean_token_accuracy": 0.7062867367267609,
|
| 836 |
+
"num_tokens": 106525361.0,
|
| 837 |
+
"step": 3750
|
| 838 |
+
},
|
| 839 |
+
{
|
| 840 |
+
"entropy": 1.4218295288085938,
|
| 841 |
+
"epoch": 1.2925170068027212,
|
| 842 |
+
"grad_norm": 0.8984375,
|
| 843 |
+
"learning_rate": 1.4041924862942744e-05,
|
| 844 |
+
"loss": 1.1394439697265626,
|
| 845 |
+
"mean_token_accuracy": 0.7095408570766449,
|
| 846 |
+
"num_tokens": 107938288.0,
|
| 847 |
+
"step": 3800
|
| 848 |
+
},
|
| 849 |
+
{
|
| 850 |
+
"entropy": 1.425144808292389,
|
| 851 |
+
"epoch": 1.3095238095238095,
|
| 852 |
+
"grad_norm": 1.1875,
|
| 853 |
+
"learning_rate": 1.3860169782658863e-05,
|
| 854 |
+
"loss": 1.1289046478271485,
|
| 855 |
+
"mean_token_accuracy": 0.7117006468772888,
|
| 856 |
+
"num_tokens": 109343324.0,
|
| 857 |
+
"step": 3850
|
| 858 |
+
},
|
| 859 |
+
{
|
| 860 |
+
"entropy": 1.4213409280776979,
|
| 861 |
+
"epoch": 1.3265306122448979,
|
| 862 |
+
"grad_norm": 1.0625,
|
| 863 |
+
"learning_rate": 1.3676903577381231e-05,
|
| 864 |
+
"loss": 1.124376220703125,
|
| 865 |
+
"mean_token_accuracy": 0.7127799916267396,
|
| 866 |
+
"num_tokens": 110773342.0,
|
| 867 |
+
"step": 3900
|
| 868 |
+
},
|
| 869 |
+
{
|
| 870 |
+
"entropy": 1.430138614177704,
|
| 871 |
+
"epoch": 1.3435374149659864,
|
| 872 |
+
"grad_norm": 1.03125,
|
| 873 |
+
"learning_rate": 1.3492197989587221e-05,
|
| 874 |
+
"loss": 1.1303221893310547,
|
| 875 |
+
"mean_token_accuracy": 0.7113751316070557,
|
| 876 |
+
"num_tokens": 112223937.0,
|
| 877 |
+
"step": 3950
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"entropy": 1.421439278125763,
|
| 881 |
+
"epoch": 1.3605442176870748,
|
| 882 |
+
"grad_norm": 1.0390625,
|
| 883 |
+
"learning_rate": 1.3306125325223433e-05,
|
| 884 |
+
"loss": 1.128666305541992,
|
| 885 |
+
"mean_token_accuracy": 0.711964373588562,
|
| 886 |
+
"num_tokens": 113638897.0,
|
| 887 |
+
"step": 4000
|
| 888 |
+
},
|
| 889 |
+
{
|
| 890 |
+
"epoch": 1.3605442176870748,
|
| 891 |
+
"eval_entropy": 1.4151852511558212,
|
| 892 |
+
"eval_loss": 1.107796311378479,
|
| 893 |
+
"eval_mean_token_accuracy": 0.7166504459220822,
|
| 894 |
+
"eval_num_tokens": 113638897.0,
|
| 895 |
+
"eval_runtime": 8.7246,
|
| 896 |
+
"eval_samples_per_second": 108.658,
|
| 897 |
+
"eval_steps_per_second": 13.64,
|
| 898 |
+
"step": 4000
|
| 899 |
+
},
|
| 900 |
+
{
|
| 901 |
+
"entropy": 1.422073357105255,
|
| 902 |
+
"epoch": 1.3775510204081631,
|
| 903 |
+
"grad_norm": 1.1015625,
|
| 904 |
+
"learning_rate": 1.31187584254004e-05,
|
| 905 |
+
"loss": 1.137821807861328,
|
| 906 |
+
"mean_token_accuracy": 0.7102219009399414,
|
| 907 |
+
"num_tokens": 115030025.0,
|
| 908 |
+
"step": 4050
|
| 909 |
+
},
|
| 910 |
+
{
|
| 911 |
+
"entropy": 1.4335832858085633,
|
| 912 |
+
"epoch": 1.3945578231292517,
|
| 913 |
+
"grad_norm": 1.1171875,
|
| 914 |
+
"learning_rate": 1.2930170637877741e-05,
|
| 915 |
+
"loss": 1.1326773834228516,
|
| 916 |
+
"mean_token_accuracy": 0.7117250227928161,
|
| 917 |
+
"num_tokens": 116464440.0,
|
| 918 |
+
"step": 4100
|
| 919 |
+
},
|
| 920 |
+
{
|
| 921 |
+
"entropy": 1.432119870185852,
|
| 922 |
+
"epoch": 1.4115646258503403,
|
| 923 |
+
"grad_norm": 1.0546875,
|
| 924 |
+
"learning_rate": 1.2740435788351031e-05,
|
| 925 |
+
"loss": 1.144314727783203,
|
| 926 |
+
"mean_token_accuracy": 0.7074739003181457,
|
| 927 |
+
"num_tokens": 117895570.0,
|
| 928 |
+
"step": 4150
|
| 929 |
+
},
|
| 930 |
+
{
|
| 931 |
+
"entropy": 1.4382991909980773,
|
| 932 |
+
"epoch": 1.4285714285714286,
|
| 933 |
+
"grad_norm": 1.0546875,
|
| 934 |
+
"learning_rate": 1.254962815155155e-05,
|
| 935 |
+
"loss": 1.1475949859619141,
|
| 936 |
+
"mean_token_accuracy": 0.7079149842262268,
|
| 937 |
+
"num_tokens": 119315574.0,
|
| 938 |
+
"step": 4200
|
| 939 |
+
},
|
| 940 |
+
{
|
| 941 |
+
"entropy": 1.3948691391944885,
|
| 942 |
+
"epoch": 1.445578231292517,
|
| 943 |
+
"grad_norm": 0.99609375,
|
| 944 |
+
"learning_rate": 1.2357822422170266e-05,
|
| 945 |
+
"loss": 1.1130772399902344,
|
| 946 |
+
"mean_token_accuracy": 0.7150168895721436,
|
| 947 |
+
"num_tokens": 120717868.0,
|
| 948 |
+
"step": 4250
|
| 949 |
+
},
|
| 950 |
+
{
|
| 951 |
+
"entropy": 1.4280868649482727,
|
| 952 |
+
"epoch": 1.4625850340136055,
|
| 953 |
+
"grad_norm": 1.015625,
|
| 954 |
+
"learning_rate": 1.2165093685617377e-05,
|
| 955 |
+
"loss": 1.1200074005126952,
|
| 956 |
+
"mean_token_accuracy": 0.7135682916641235,
|
| 957 |
+
"num_tokens": 122146924.0,
|
| 958 |
+
"step": 4300
|
| 959 |
+
},
|
| 960 |
+
{
|
| 961 |
+
"entropy": 1.4316187024116516,
|
| 962 |
+
"epoch": 1.4795918367346939,
|
| 963 |
+
"grad_norm": 1.0390625,
|
| 964 |
+
"learning_rate": 1.1971517388628972e-05,
|
| 965 |
+
"loss": 1.1323360443115233,
|
| 966 |
+
"mean_token_accuracy": 0.7125037395954132,
|
| 967 |
+
"num_tokens": 123574594.0,
|
| 968 |
+
"step": 4350
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"entropy": 1.4233829760551453,
|
| 972 |
+
"epoch": 1.4965986394557822,
|
| 973 |
+
"grad_norm": 1.0625,
|
| 974 |
+
"learning_rate": 1.1777169309732178e-05,
|
| 975 |
+
"loss": 1.144235610961914,
|
| 976 |
+
"mean_token_accuracy": 0.707960147857666,
|
| 977 |
+
"num_tokens": 124985723.0,
|
| 978 |
+
"step": 4400
|
| 979 |
+
},
|
| 980 |
+
{
|
| 981 |
+
"entropy": 1.446813395023346,
|
| 982 |
+
"epoch": 1.5136054421768708,
|
| 983 |
+
"grad_norm": 0.9765625,
|
| 984 |
+
"learning_rate": 1.1582125529580437e-05,
|
| 985 |
+
"loss": 1.1620960235595703,
|
| 986 |
+
"mean_token_accuracy": 0.7044985663890838,
|
| 987 |
+
"num_tokens": 126444415.0,
|
| 988 |
+
"step": 4450
|
| 989 |
+
},
|
| 990 |
+
{
|
| 991 |
+
"entropy": 1.4292146420478822,
|
| 992 |
+
"epoch": 1.5306122448979593,
|
| 993 |
+
"grad_norm": 1.1328125,
|
| 994 |
+
"learning_rate": 1.1386462401170562e-05,
|
| 995 |
+
"loss": 1.1338434600830078,
|
| 996 |
+
"mean_token_accuracy": 0.7091768038272858,
|
| 997 |
+
"num_tokens": 127870485.0,
|
| 998 |
+
"step": 4500
|
| 999 |
+
},
|
| 1000 |
+
{
|
| 1001 |
+
"epoch": 1.5306122448979593,
|
| 1002 |
+
"eval_entropy": 1.4135069837089347,
|
| 1003 |
+
"eval_loss": 1.106415033340454,
|
| 1004 |
+
"eval_mean_token_accuracy": 0.7169664757592338,
|
| 1005 |
+
"eval_num_tokens": 127870485.0,
|
| 1006 |
+
"eval_runtime": 8.6098,
|
| 1007 |
+
"eval_samples_per_second": 110.107,
|
| 1008 |
+
"eval_steps_per_second": 13.821,
|
| 1009 |
+
"step": 4500
|
| 1010 |
+
},
|
| 1011 |
+
{
|
| 1012 |
+
"entropy": 1.412301104068756,
|
| 1013 |
+
"epoch": 1.5476190476190477,
|
| 1014 |
+
"grad_norm": 1.1171875,
|
| 1015 |
+
"learning_rate": 1.1190256519953097e-05,
|
| 1016 |
+
"loss": 1.1280068206787108,
|
| 1017 |
+
"mean_token_accuracy": 0.7108085489273072,
|
| 1018 |
+
"num_tokens": 129318094.0,
|
| 1019 |
+
"step": 4550
|
| 1020 |
+
},
|
| 1021 |
+
{
|
| 1022 |
+
"entropy": 1.4223233222961427,
|
| 1023 |
+
"epoch": 1.564625850340136,
|
| 1024 |
+
"grad_norm": 1.1484375,
|
| 1025 |
+
"learning_rate": 1.0993584693847827e-05,
|
| 1026 |
+
"loss": 1.1158720397949218,
|
| 1027 |
+
"mean_token_accuracy": 0.7139767467975616,
|
| 1028 |
+
"num_tokens": 130726607.0,
|
| 1029 |
+
"step": 4600
|
| 1030 |
+
},
|
| 1031 |
+
{
|
| 1032 |
+
"entropy": 1.42511239528656,
|
| 1033 |
+
"epoch": 1.5816326530612246,
|
| 1034 |
+
"grad_norm": 1.1171875,
|
| 1035 |
+
"learning_rate": 1.0796523913176092e-05,
|
| 1036 |
+
"loss": 1.1147970581054687,
|
| 1037 |
+
"mean_token_accuracy": 0.7152622473239899,
|
| 1038 |
+
"num_tokens": 132145387.0,
|
| 1039 |
+
"step": 4650
|
| 1040 |
+
},
|
| 1041 |
+
{
|
| 1042 |
+
"entropy": 1.4350685811042785,
|
| 1043 |
+
"epoch": 1.598639455782313,
|
| 1044 |
+
"grad_norm": 1.0625,
|
| 1045 |
+
"learning_rate": 1.0599151320521703e-05,
|
| 1046 |
+
"loss": 1.1316419982910155,
|
| 1047 |
+
"mean_token_accuracy": 0.7108878707885742,
|
| 1048 |
+
"num_tokens": 133560830.0,
|
| 1049 |
+
"step": 4700
|
| 1050 |
+
},
|
| 1051 |
+
{
|
| 1052 |
+
"entropy": 1.4368190813064574,
|
| 1053 |
+
"epoch": 1.6156462585034013,
|
| 1054 |
+
"grad_norm": 1.0703125,
|
| 1055 |
+
"learning_rate": 1.040154418053223e-05,
|
| 1056 |
+
"loss": 1.1656169128417968,
|
| 1057 |
+
"mean_token_accuracy": 0.7034963488578796,
|
| 1058 |
+
"num_tokens": 134992182.0,
|
| 1059 |
+
"step": 4750
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"entropy": 1.4219933152198792,
|
| 1063 |
+
"epoch": 1.6326530612244898,
|
| 1064 |
+
"grad_norm": 1.0234375,
|
| 1065 |
+
"learning_rate": 1.0203779849672546e-05,
|
| 1066 |
+
"loss": 1.1332920837402343,
|
| 1067 |
+
"mean_token_accuracy": 0.7100002634525299,
|
| 1068 |
+
"num_tokens": 136426899.0,
|
| 1069 |
+
"step": 4800
|
| 1070 |
+
},
|
| 1071 |
+
{
|
| 1072 |
+
"entropy": 1.4121305012702943,
|
| 1073 |
+
"epoch": 1.6496598639455784,
|
| 1074 |
+
"grad_norm": 1.1484375,
|
| 1075 |
+
"learning_rate": 1.0005935745942395e-05,
|
| 1076 |
+
"loss": 1.120484619140625,
|
| 1077 |
+
"mean_token_accuracy": 0.7137669050693511,
|
| 1078 |
+
"num_tokens": 137821124.0,
|
| 1079 |
+
"step": 4850
|
| 1080 |
+
},
|
| 1081 |
+
{
|
| 1082 |
+
"entropy": 1.4223187255859375,
|
| 1083 |
+
"epoch": 1.6666666666666665,
|
| 1084 |
+
"grad_norm": 1.21875,
|
| 1085 |
+
"learning_rate": 9.808089318569872e-06,
|
| 1086 |
+
"loss": 1.13340087890625,
|
| 1087 |
+
"mean_token_accuracy": 0.7105760216712952,
|
| 1088 |
+
"num_tokens": 139238831.0,
|
| 1089 |
+
"step": 4900
|
| 1090 |
+
},
|
| 1091 |
+
{
|
| 1092 |
+
"entropy": 1.4339079999923705,
|
| 1093 |
+
"epoch": 1.683673469387755,
|
| 1094 |
+
"grad_norm": 1.1171875,
|
| 1095 |
+
"learning_rate": 9.610318017692717e-06,
|
| 1096 |
+
"loss": 1.131591567993164,
|
| 1097 |
+
"mean_token_accuracy": 0.7117097544670105,
|
| 1098 |
+
"num_tokens": 140670083.0,
|
| 1099 |
+
"step": 4950
|
| 1100 |
+
},
|
| 1101 |
+
{
|
| 1102 |
+
"entropy": 1.42487220287323,
|
| 1103 |
+
"epoch": 1.7006802721088436,
|
| 1104 |
+
"grad_norm": 1.0,
|
| 1105 |
+
"learning_rate": 9.412699264039187e-06,
|
| 1106 |
+
"loss": 1.1361915588378906,
|
| 1107 |
+
"mean_token_accuracy": 0.709000414609909,
|
| 1108 |
+
"num_tokens": 142109640.0,
|
| 1109 |
+
"step": 5000
|
| 1110 |
+
},
|
| 1111 |
+
{
|
| 1112 |
+
"epoch": 1.7006802721088436,
|
| 1113 |
+
"eval_entropy": 1.4130272725049187,
|
| 1114 |
+
"eval_loss": 1.105644941329956,
|
| 1115 |
+
"eval_mean_token_accuracy": 0.7170687387971317,
|
| 1116 |
+
"eval_num_tokens": 142109640.0,
|
| 1117 |
+
"eval_runtime": 9.1754,
|
| 1118 |
+
"eval_samples_per_second": 103.32,
|
| 1119 |
+
"eval_steps_per_second": 12.97,
|
| 1120 |
+
"step": 5000
|
| 1121 |
+
},
|
| 1122 |
+
{
|
| 1123 |
+
"entropy": 1.43812607049942,
|
| 1124 |
+
"epoch": 1.717687074829932,
|
| 1125 |
+
"grad_norm": 1.109375,
|
| 1126 |
+
"learning_rate": 9.215310418620513e-06,
|
| 1127 |
+
"loss": 1.1505773162841797,
|
| 1128 |
+
"mean_token_accuracy": 0.7071945500373841,
|
| 1129 |
+
"num_tokens": 143529976.0,
|
| 1130 |
+
"step": 5050
|
| 1131 |
+
},
|
| 1132 |
+
{
|
| 1133 |
+
"entropy": 1.4412685680389403,
|
| 1134 |
+
"epoch": 1.7346938775510203,
|
| 1135 |
+
"grad_norm": 1.2890625,
|
| 1136 |
+
"learning_rate": 9.018228752446671e-06,
|
| 1137 |
+
"loss": 1.1481983947753907,
|
| 1138 |
+
"mean_token_accuracy": 0.7087244749069214,
|
| 1139 |
+
"num_tokens": 144937854.0,
|
| 1140 |
+
"step": 5100
|
| 1141 |
+
},
|
| 1142 |
+
{
|
| 1143 |
+
"entropy": 1.4139077353477478,
|
| 1144 |
+
"epoch": 1.751700680272109,
|
| 1145 |
+
"grad_norm": 1.046875,
|
| 1146 |
+
"learning_rate": 8.821531416277427e-06,
|
| 1147 |
+
"loss": 1.1202762603759766,
|
| 1148 |
+
"mean_token_accuracy": 0.7141082406044006,
|
| 1149 |
+
"num_tokens": 146358885.0,
|
| 1150 |
+
"step": 5150
|
| 1151 |
+
},
|
| 1152 |
+
{
|
| 1153 |
+
"entropy": 1.4265229201316834,
|
| 1154 |
+
"epoch": 1.7687074829931972,
|
| 1155 |
+
"grad_norm": 1.0390625,
|
| 1156 |
+
"learning_rate": 8.625295410420451e-06,
|
| 1157 |
+
"loss": 1.1334566497802734,
|
| 1158 |
+
"mean_token_accuracy": 0.7119065308570862,
|
| 1159 |
+
"num_tokens": 147797164.0,
|
| 1160 |
+
"step": 5200
|
| 1161 |
+
},
|
| 1162 |
+
{
|
| 1163 |
+
"entropy": 1.4280506682395935,
|
| 1164 |
+
"epoch": 1.7857142857142856,
|
| 1165 |
+
"grad_norm": 1.125,
|
| 1166 |
+
"learning_rate": 8.429597554588291e-06,
|
| 1167 |
+
"loss": 1.1369068908691407,
|
| 1168 |
+
"mean_token_accuracy": 0.7097280561923981,
|
| 1169 |
+
"num_tokens": 149201547.0,
|
| 1170 |
+
"step": 5250
|
| 1171 |
+
},
|
| 1172 |
+
{
|
| 1173 |
+
"entropy": 1.4155354285240174,
|
| 1174 |
+
"epoch": 1.8027210884353742,
|
| 1175 |
+
"grad_norm": 1.0390625,
|
| 1176 |
+
"learning_rate": 8.234514457826101e-06,
|
| 1177 |
+
"loss": 1.1342711639404297,
|
| 1178 |
+
"mean_token_accuracy": 0.7102336215972901,
|
| 1179 |
+
"num_tokens": 150625894.0,
|
| 1180 |
+
"step": 5300
|
| 1181 |
+
},
|
| 1182 |
+
{
|
| 1183 |
+
"entropy": 1.4110008382797241,
|
| 1184 |
+
"epoch": 1.8197278911564627,
|
| 1185 |
+
"grad_norm": 1.1640625,
|
| 1186 |
+
"learning_rate": 8.040122488521802e-06,
|
| 1187 |
+
"loss": 1.1284471893310546,
|
| 1188 |
+
"mean_token_accuracy": 0.7116459333896636,
|
| 1189 |
+
"num_tokens": 152068477.0,
|
| 1190 |
+
"step": 5350
|
| 1191 |
+
},
|
| 1192 |
+
{
|
| 1193 |
+
"entropy": 1.432766444683075,
|
| 1194 |
+
"epoch": 1.836734693877551,
|
| 1195 |
+
"grad_norm": 1.109375,
|
| 1196 |
+
"learning_rate": 7.84649774451044e-06,
|
| 1197 |
+
"loss": 1.1442298126220702,
|
| 1198 |
+
"mean_token_accuracy": 0.7084469604492187,
|
| 1199 |
+
"num_tokens": 153452805.0,
|
| 1200 |
+
"step": 5400
|
| 1201 |
+
},
|
| 1202 |
+
{
|
| 1203 |
+
"entropy": 1.4228289747238159,
|
| 1204 |
+
"epoch": 1.8537414965986394,
|
| 1205 |
+
"grad_norm": 1.0,
|
| 1206 |
+
"learning_rate": 7.653716023284498e-06,
|
| 1207 |
+
"loss": 1.1224955749511718,
|
| 1208 |
+
"mean_token_accuracy": 0.7124292123317718,
|
| 1209 |
+
"num_tokens": 154873137.0,
|
| 1210 |
+
"step": 5450
|
| 1211 |
+
},
|
| 1212 |
+
{
|
| 1213 |
+
"entropy": 1.4096328234672546,
|
| 1214 |
+
"epoch": 1.870748299319728,
|
| 1215 |
+
"grad_norm": 1.109375,
|
| 1216 |
+
"learning_rate": 7.46185279232176e-06,
|
| 1217 |
+
"loss": 1.1197357940673829,
|
| 1218 |
+
"mean_token_accuracy": 0.7131345391273498,
|
| 1219 |
+
"num_tokens": 156324700.0,
|
| 1220 |
+
"step": 5500
|
| 1221 |
+
},
|
| 1222 |
+
{
|
| 1223 |
+
"epoch": 1.870748299319728,
|
| 1224 |
+
"eval_entropy": 1.4128466764418017,
|
| 1225 |
+
"eval_loss": 1.1052268743515015,
|
| 1226 |
+
"eval_mean_token_accuracy": 0.7171510763528968,
|
| 1227 |
+
"eval_num_tokens": 156324700.0,
|
| 1228 |
+
"eval_runtime": 9.0481,
|
| 1229 |
+
"eval_samples_per_second": 104.773,
|
| 1230 |
+
"eval_steps_per_second": 13.152,
|
| 1231 |
+
"step": 5500
|
| 1232 |
+
},
|
| 1233 |
+
{
|
| 1234 |
+
"entropy": 1.4428221464157105,
|
| 1235 |
+
"epoch": 1.8877551020408163,
|
| 1236 |
+
"grad_norm": 0.96484375,
|
| 1237 |
+
"learning_rate": 7.27098315954234e-06,
|
| 1238 |
+
"loss": 1.1470510101318359,
|
| 1239 |
+
"mean_token_accuracy": 0.7073501777648926,
|
| 1240 |
+
"num_tokens": 157737046.0,
|
| 1241 |
+
"step": 5550
|
| 1242 |
+
},
|
| 1243 |
+
{
|
| 1244 |
+
"entropy": 1.4279747819900512,
|
| 1245 |
+
"epoch": 1.9047619047619047,
|
| 1246 |
+
"grad_norm": 1.0390625,
|
| 1247 |
+
"learning_rate": 7.081181843906513e-06,
|
| 1248 |
+
"loss": 1.1433404541015626,
|
| 1249 |
+
"mean_token_accuracy": 0.7082498788833618,
|
| 1250 |
+
"num_tokens": 159172992.0,
|
| 1251 |
+
"step": 5600
|
| 1252 |
+
},
|
| 1253 |
+
{
|
| 1254 |
+
"entropy": 1.4385703134536743,
|
| 1255 |
+
"epoch": 1.9217687074829932,
|
| 1256 |
+
"grad_norm": 1.1328125,
|
| 1257 |
+
"learning_rate": 6.892523146164782e-06,
|
| 1258 |
+
"loss": 1.1372301483154297,
|
| 1259 |
+
"mean_token_accuracy": 0.7099990272521972,
|
| 1260 |
+
"num_tokens": 160577822.0,
|
| 1261 |
+
"step": 5650
|
| 1262 |
+
},
|
| 1263 |
+
{
|
| 1264 |
+
"entropy": 1.4102177619934082,
|
| 1265 |
+
"epoch": 1.9387755102040818,
|
| 1266 |
+
"grad_norm": 1.0625,
|
| 1267 |
+
"learning_rate": 6.705080919771655e-06,
|
| 1268 |
+
"loss": 1.1329367065429687,
|
| 1269 |
+
"mean_token_accuracy": 0.7119775557518006,
|
| 1270 |
+
"num_tokens": 161991010.0,
|
| 1271 |
+
"step": 5700
|
| 1272 |
+
},
|
| 1273 |
+
{
|
| 1274 |
+
"entropy": 1.4429755330085754,
|
| 1275 |
+
"epoch": 1.95578231292517,
|
| 1276 |
+
"grad_norm": 1.125,
|
| 1277 |
+
"learning_rate": 6.51892854197454e-06,
|
| 1278 |
+
"loss": 1.1424610900878907,
|
| 1279 |
+
"mean_token_accuracy": 0.7083716082572937,
|
| 1280 |
+
"num_tokens": 163405551.0,
|
| 1281 |
+
"step": 5750
|
| 1282 |
+
},
|
| 1283 |
+
{
|
| 1284 |
+
"entropy": 1.4363023614883423,
|
| 1285 |
+
"epoch": 1.9727891156462585,
|
| 1286 |
+
"grad_norm": 1.109375,
|
| 1287 |
+
"learning_rate": 6.334138885089061e-06,
|
| 1288 |
+
"loss": 1.142376937866211,
|
| 1289 |
+
"mean_token_accuracy": 0.709513326883316,
|
| 1290 |
+
"num_tokens": 164828847.0,
|
| 1291 |
+
"step": 5800
|
| 1292 |
+
},
|
| 1293 |
+
{
|
| 1294 |
+
"entropy": 1.435585732460022,
|
| 1295 |
+
"epoch": 1.989795918367347,
|
| 1296 |
+
"grad_norm": 1.0234375,
|
| 1297 |
+
"learning_rate": 6.150784287972005e-06,
|
| 1298 |
+
"loss": 1.157130126953125,
|
| 1299 |
+
"mean_token_accuracy": 0.7060146403312683,
|
| 1300 |
+
"num_tokens": 166264927.0,
|
| 1301 |
+
"step": 5850
|
| 1302 |
+
}
|
| 1303 |
+
],
|
| 1304 |
+
"logging_steps": 50,
|
| 1305 |
+
"max_steps": 8820,
|
| 1306 |
+
"num_input_tokens_seen": 0,
|
| 1307 |
+
"num_train_epochs": 3,
|
| 1308 |
+
"save_steps": 500,
|
| 1309 |
+
"stateful_callbacks": {
|
| 1310 |
+
"TrainerControl": {
|
| 1311 |
+
"args": {
|
| 1312 |
+
"should_epoch_stop": false,
|
| 1313 |
+
"should_evaluate": false,
|
| 1314 |
+
"should_log": false,
|
| 1315 |
+
"should_save": true,
|
| 1316 |
+
"should_training_stop": false
|
| 1317 |
+
},
|
| 1318 |
+
"attributes": {}
|
| 1319 |
+
}
|
| 1320 |
+
},
|
| 1321 |
+
"total_flos": 9.18336854621376e+17,
|
| 1322 |
+
"train_batch_size": 32,
|
| 1323 |
+
"trial_name": null,
|
| 1324 |
+
"trial_params": null
|
| 1325 |
+
}
|
checkpoint-5880/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2309a8eab46e70d77a5e11e1a0025a30065cc758ddf39c81cb9137e50f03b60b
|
| 3 |
+
size 5649
|
checkpoint-8820/config.json
ADDED
|
@@ -0,0 +1,32 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"LlamaForCausalLM"
|
| 4 |
+
],
|
| 5 |
+
"attention_bias": false,
|
| 6 |
+
"attention_dropout": 0.0,
|
| 7 |
+
"bos_token_id": 0,
|
| 8 |
+
"dtype": "bfloat16",
|
| 9 |
+
"eos_token_id": 0,
|
| 10 |
+
"head_dim": 64,
|
| 11 |
+
"hidden_act": "silu",
|
| 12 |
+
"hidden_size": 1280,
|
| 13 |
+
"initializer_range": 0.02,
|
| 14 |
+
"intermediate_size": 2816,
|
| 15 |
+
"max_position_embeddings": 2048,
|
| 16 |
+
"mlp_bias": false,
|
| 17 |
+
"model_type": "llama",
|
| 18 |
+
"num_attention_heads": 20,
|
| 19 |
+
"num_hidden_layers": 24,
|
| 20 |
+
"num_key_value_heads": 10,
|
| 21 |
+
"pad_token_id": 1,
|
| 22 |
+
"pretraining_tp": 1,
|
| 23 |
+
"rms_norm_eps": 1e-05,
|
| 24 |
+
"rope_parameters": {
|
| 25 |
+
"rope_theta": 10000.0,
|
| 26 |
+
"rope_type": "default"
|
| 27 |
+
},
|
| 28 |
+
"tie_word_embeddings": false,
|
| 29 |
+
"transformers_version": "5.3.0",
|
| 30 |
+
"use_cache": false,
|
| 31 |
+
"vocab_size": 50304
|
| 32 |
+
}
|
checkpoint-8820/generation_config.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"bos_token_id": 0,
|
| 4 |
+
"eos_token_id": [
|
| 5 |
+
0
|
| 6 |
+
],
|
| 7 |
+
"pad_token_id": 1,
|
| 8 |
+
"transformers_version": "5.3.0"
|
| 9 |
+
}
|
checkpoint-8820/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b74230cccba97ff7a7d718cc33463a9830d024351a5b2eb867a0285f2fbe8ea7
|
| 3 |
+
size 1012681552
|
checkpoint-8820/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:dcdc6f7a27602596ffda38a989c9c8c131d5b8c2b7246e94623f13148cc0b25e
|
| 3 |
+
size 2025501387
|
checkpoint-8820/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ed3753ab7977739b8eda494dd72defae5750f7283141b11a8f562160ba4c1a23
|
| 3 |
+
size 14645
|
checkpoint-8820/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:fac1c5f2e36054d858f26b63b395bd7c3020ee27b5d55487e4f05fa68e66f48c
|
| 3 |
+
size 1465
|
checkpoint-8820/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
checkpoint-8820/tokenizer_config.json
ADDED
|
@@ -0,0 +1,21 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_prefix_space": false,
|
| 3 |
+
"backend": "tokenizers",
|
| 4 |
+
"bos_token": "<|endoftext|>",
|
| 5 |
+
"eos_token": "<|endoftext|>",
|
| 6 |
+
"errors": "replace",
|
| 7 |
+
"extra_special_tokens": [
|
| 8 |
+
"<|system|>",
|
| 9 |
+
"<|user|>",
|
| 10 |
+
"<|assistant|>",
|
| 11 |
+
"<|tool_call|>",
|
| 12 |
+
"<|tool_result|>",
|
| 13 |
+
"<|end|>"
|
| 14 |
+
],
|
| 15 |
+
"is_local": true,
|
| 16 |
+
"model_max_length": 1000000000000000019884624838656,
|
| 17 |
+
"pad_token": "<|padding|>",
|
| 18 |
+
"tokenizer_class": "GPTNeoXTokenizer",
|
| 19 |
+
"trim_offsets": true,
|
| 20 |
+
"unk_token": "<|endoftext|>"
|
| 21 |
+
}
|
checkpoint-8820/trainer_state.json
ADDED
|
@@ -0,0 +1,1981 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": null,
|
| 3 |
+
"best_metric": null,
|
| 4 |
+
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 3.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 8820,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"entropy": 1.7698939561843872,
|
| 14 |
+
"epoch": 0.017006802721088437,
|
| 15 |
+
"grad_norm": 3.796875,
|
| 16 |
+
"learning_rate": 1.112372304199773e-06,
|
| 17 |
+
"loss": 1.5452156066894531,
|
| 18 |
+
"mean_token_accuracy": 0.6630511581897736,
|
| 19 |
+
"num_tokens": 1416855.0,
|
| 20 |
+
"step": 50
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"entropy": 1.765206663608551,
|
| 24 |
+
"epoch": 0.034013605442176874,
|
| 25 |
+
"grad_norm": 4.6875,
|
| 26 |
+
"learning_rate": 2.2474460839954596e-06,
|
| 27 |
+
"loss": 1.5233023071289062,
|
| 28 |
+
"mean_token_accuracy": 0.6625268197059632,
|
| 29 |
+
"num_tokens": 2861955.0,
|
| 30 |
+
"step": 100
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"entropy": 1.7640055298805237,
|
| 34 |
+
"epoch": 0.05102040816326531,
|
| 35 |
+
"grad_norm": 2.984375,
|
| 36 |
+
"learning_rate": 3.3825198637911465e-06,
|
| 37 |
+
"loss": 1.4435250854492188,
|
| 38 |
+
"mean_token_accuracy": 0.6652392935752869,
|
| 39 |
+
"num_tokens": 4279472.0,
|
| 40 |
+
"step": 150
|
| 41 |
+
},
|
| 42 |
+
{
|
| 43 |
+
"entropy": 1.7068938755989074,
|
| 44 |
+
"epoch": 0.06802721088435375,
|
| 45 |
+
"grad_norm": 1.3828125,
|
| 46 |
+
"learning_rate": 4.517593643586833e-06,
|
| 47 |
+
"loss": 1.3378631591796875,
|
| 48 |
+
"mean_token_accuracy": 0.673524329662323,
|
| 49 |
+
"num_tokens": 5707775.0,
|
| 50 |
+
"step": 200
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"entropy": 1.6458502221107483,
|
| 54 |
+
"epoch": 0.08503401360544217,
|
| 55 |
+
"grad_norm": 1.1015625,
|
| 56 |
+
"learning_rate": 5.652667423382521e-06,
|
| 57 |
+
"loss": 1.2989051818847657,
|
| 58 |
+
"mean_token_accuracy": 0.6784149742126465,
|
| 59 |
+
"num_tokens": 7118906.0,
|
| 60 |
+
"step": 250
|
| 61 |
+
},
|
| 62 |
+
{
|
| 63 |
+
"entropy": 1.6340991377830505,
|
| 64 |
+
"epoch": 0.10204081632653061,
|
| 65 |
+
"grad_norm": 1.21875,
|
| 66 |
+
"learning_rate": 6.787741203178207e-06,
|
| 67 |
+
"loss": 1.2801470947265625,
|
| 68 |
+
"mean_token_accuracy": 0.6825140416622162,
|
| 69 |
+
"num_tokens": 8529671.0,
|
| 70 |
+
"step": 300
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"entropy": 1.599980125427246,
|
| 74 |
+
"epoch": 0.11904761904761904,
|
| 75 |
+
"grad_norm": 1.1640625,
|
| 76 |
+
"learning_rate": 7.922814982973895e-06,
|
| 77 |
+
"loss": 1.2615231323242186,
|
| 78 |
+
"mean_token_accuracy": 0.6857551598548889,
|
| 79 |
+
"num_tokens": 9973401.0,
|
| 80 |
+
"step": 350
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"entropy": 1.570529818534851,
|
| 84 |
+
"epoch": 0.1360544217687075,
|
| 85 |
+
"grad_norm": 1.078125,
|
| 86 |
+
"learning_rate": 9.057888762769582e-06,
|
| 87 |
+
"loss": 1.2502473449707032,
|
| 88 |
+
"mean_token_accuracy": 0.6881939315795899,
|
| 89 |
+
"num_tokens": 11391611.0,
|
| 90 |
+
"step": 400
|
| 91 |
+
},
|
| 92 |
+
{
|
| 93 |
+
"entropy": 1.5739640021324157,
|
| 94 |
+
"epoch": 0.15306122448979592,
|
| 95 |
+
"grad_norm": 1.1640625,
|
| 96 |
+
"learning_rate": 1.0192962542565269e-05,
|
| 97 |
+
"loss": 1.2381304168701173,
|
| 98 |
+
"mean_token_accuracy": 0.6904172372817993,
|
| 99 |
+
"num_tokens": 12818742.0,
|
| 100 |
+
"step": 450
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"entropy": 1.542729480266571,
|
| 104 |
+
"epoch": 0.17006802721088435,
|
| 105 |
+
"grad_norm": 1.2421875,
|
| 106 |
+
"learning_rate": 1.1328036322360954e-05,
|
| 107 |
+
"loss": 1.2234064483642577,
|
| 108 |
+
"mean_token_accuracy": 0.6928502190113067,
|
| 109 |
+
"num_tokens": 14230694.0,
|
| 110 |
+
"step": 500
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.17006802721088435,
|
| 114 |
+
"eval_entropy": 1.5409097210699771,
|
| 115 |
+
"eval_loss": 1.1968441009521484,
|
| 116 |
+
"eval_mean_token_accuracy": 0.6989319805337602,
|
| 117 |
+
"eval_num_tokens": 14230694.0,
|
| 118 |
+
"eval_runtime": 9.3471,
|
| 119 |
+
"eval_samples_per_second": 101.422,
|
| 120 |
+
"eval_steps_per_second": 12.731,
|
| 121 |
+
"step": 500
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"entropy": 1.5507296323776245,
|
| 125 |
+
"epoch": 0.1870748299319728,
|
| 126 |
+
"grad_norm": 1.2265625,
|
| 127 |
+
"learning_rate": 1.246311010215664e-05,
|
| 128 |
+
"loss": 1.2286612701416015,
|
| 129 |
+
"mean_token_accuracy": 0.6924433386325837,
|
| 130 |
+
"num_tokens": 15656461.0,
|
| 131 |
+
"step": 550
|
| 132 |
+
},
|
| 133 |
+
{
|
| 134 |
+
"entropy": 1.5728167843818666,
|
| 135 |
+
"epoch": 0.20408163265306123,
|
| 136 |
+
"grad_norm": 1.21875,
|
| 137 |
+
"learning_rate": 1.3598183881952328e-05,
|
| 138 |
+
"loss": 1.2478004455566407,
|
| 139 |
+
"mean_token_accuracy": 0.6871815168857575,
|
| 140 |
+
"num_tokens": 17079797.0,
|
| 141 |
+
"step": 600
|
| 142 |
+
},
|
| 143 |
+
{
|
| 144 |
+
"entropy": 1.5617320823669434,
|
| 145 |
+
"epoch": 0.22108843537414966,
|
| 146 |
+
"grad_norm": 1.125,
|
| 147 |
+
"learning_rate": 1.4733257661748014e-05,
|
| 148 |
+
"loss": 1.2422699737548828,
|
| 149 |
+
"mean_token_accuracy": 0.6885153651237488,
|
| 150 |
+
"num_tokens": 18519114.0,
|
| 151 |
+
"step": 650
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"entropy": 1.54343731880188,
|
| 155 |
+
"epoch": 0.23809523809523808,
|
| 156 |
+
"grad_norm": 1.1484375,
|
| 157 |
+
"learning_rate": 1.58683314415437e-05,
|
| 158 |
+
"loss": 1.2279574584960937,
|
| 159 |
+
"mean_token_accuracy": 0.6913824212551117,
|
| 160 |
+
"num_tokens": 19930551.0,
|
| 161 |
+
"step": 700
|
| 162 |
+
},
|
| 163 |
+
{
|
| 164 |
+
"entropy": 1.524279227256775,
|
| 165 |
+
"epoch": 0.25510204081632654,
|
| 166 |
+
"grad_norm": 1.140625,
|
| 167 |
+
"learning_rate": 1.700340522133939e-05,
|
| 168 |
+
"loss": 1.2193751525878906,
|
| 169 |
+
"mean_token_accuracy": 0.6942444491386414,
|
| 170 |
+
"num_tokens": 21345785.0,
|
| 171 |
+
"step": 750
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"entropy": 1.527976930141449,
|
| 175 |
+
"epoch": 0.272108843537415,
|
| 176 |
+
"grad_norm": 1.1171875,
|
| 177 |
+
"learning_rate": 1.8138479001135073e-05,
|
| 178 |
+
"loss": 1.212171859741211,
|
| 179 |
+
"mean_token_accuracy": 0.6960112357139587,
|
| 180 |
+
"num_tokens": 22754238.0,
|
| 181 |
+
"step": 800
|
| 182 |
+
},
|
| 183 |
+
{
|
| 184 |
+
"entropy": 1.5164598178863526,
|
| 185 |
+
"epoch": 0.2891156462585034,
|
| 186 |
+
"grad_norm": 1.109375,
|
| 187 |
+
"learning_rate": 1.9273552780930764e-05,
|
| 188 |
+
"loss": 1.2044546508789062,
|
| 189 |
+
"mean_token_accuracy": 0.6967348265647888,
|
| 190 |
+
"num_tokens": 24167833.0,
|
| 191 |
+
"step": 850
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"entropy": 1.4968056178092957,
|
| 195 |
+
"epoch": 0.30612244897959184,
|
| 196 |
+
"grad_norm": 1.0234375,
|
| 197 |
+
"learning_rate": 1.9999746322867523e-05,
|
| 198 |
+
"loss": 1.2030063629150392,
|
| 199 |
+
"mean_token_accuracy": 0.6967223298549652,
|
| 200 |
+
"num_tokens": 25597839.0,
|
| 201 |
+
"step": 900
|
| 202 |
+
},
|
| 203 |
+
{
|
| 204 |
+
"entropy": 1.5187969017028808,
|
| 205 |
+
"epoch": 0.3231292517006803,
|
| 206 |
+
"grad_norm": 1.1796875,
|
| 207 |
+
"learning_rate": 1.9996379823325586e-05,
|
| 208 |
+
"loss": 1.2039017486572265,
|
| 209 |
+
"mean_token_accuracy": 0.69672048330307,
|
| 210 |
+
"num_tokens": 27039491.0,
|
| 211 |
+
"step": 950
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"entropy": 1.5000163841247558,
|
| 215 |
+
"epoch": 0.3401360544217687,
|
| 216 |
+
"grad_norm": 1.0546875,
|
| 217 |
+
"learning_rate": 1.9989100081556328e-05,
|
| 218 |
+
"loss": 1.2096540832519531,
|
| 219 |
+
"mean_token_accuracy": 0.696036319732666,
|
| 220 |
+
"num_tokens": 28462525.0,
|
| 221 |
+
"step": 1000
|
| 222 |
+
},
|
| 223 |
+
{
|
| 224 |
+
"epoch": 0.3401360544217687,
|
| 225 |
+
"eval_entropy": 1.4825329389892707,
|
| 226 |
+
"eval_loss": 1.1529995203018188,
|
| 227 |
+
"eval_mean_token_accuracy": 0.7072621358542883,
|
| 228 |
+
"eval_num_tokens": 28462525.0,
|
| 229 |
+
"eval_runtime": 9.42,
|
| 230 |
+
"eval_samples_per_second": 100.637,
|
| 231 |
+
"eval_steps_per_second": 12.633,
|
| 232 |
+
"step": 1000
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"entropy": 1.4857492208480836,
|
| 236 |
+
"epoch": 0.35714285714285715,
|
| 237 |
+
"grad_norm": 1.0625,
|
| 238 |
+
"learning_rate": 1.997790994733071e-05,
|
| 239 |
+
"loss": 1.1972512054443358,
|
| 240 |
+
"mean_token_accuracy": 0.6976107740402222,
|
| 241 |
+
"num_tokens": 29880725.0,
|
| 242 |
+
"step": 1050
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"entropy": 1.4679273128509522,
|
| 246 |
+
"epoch": 0.3741496598639456,
|
| 247 |
+
"grad_norm": 1.125,
|
| 248 |
+
"learning_rate": 1.9962813801205148e-05,
|
| 249 |
+
"loss": 1.1646170806884766,
|
| 250 |
+
"mean_token_accuracy": 0.705207599401474,
|
| 251 |
+
"num_tokens": 31288250.0,
|
| 252 |
+
"step": 1100
|
| 253 |
+
},
|
| 254 |
+
{
|
| 255 |
+
"entropy": 1.482924530506134,
|
| 256 |
+
"epoch": 0.391156462585034,
|
| 257 |
+
"grad_norm": 1.25,
|
| 258 |
+
"learning_rate": 1.9943817552806678e-05,
|
| 259 |
+
"loss": 1.1940898132324218,
|
| 260 |
+
"mean_token_accuracy": 0.6988638079166413,
|
| 261 |
+
"num_tokens": 32696424.0,
|
| 262 |
+
"step": 1150
|
| 263 |
+
},
|
| 264 |
+
{
|
| 265 |
+
"entropy": 1.4768511247634888,
|
| 266 |
+
"epoch": 0.40816326530612246,
|
| 267 |
+
"grad_norm": 1.1015625,
|
| 268 |
+
"learning_rate": 1.9920928638519545e-05,
|
| 269 |
+
"loss": 1.1782675933837892,
|
| 270 |
+
"mean_token_accuracy": 0.7024707806110382,
|
| 271 |
+
"num_tokens": 34115752.0,
|
| 272 |
+
"step": 1200
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"entropy": 1.4849171042442322,
|
| 276 |
+
"epoch": 0.42517006802721086,
|
| 277 |
+
"grad_norm": 1.1015625,
|
| 278 |
+
"learning_rate": 1.9894156018574104e-05,
|
| 279 |
+
"loss": 1.1730673217773437,
|
| 280 |
+
"mean_token_accuracy": 0.7031274008750915,
|
| 281 |
+
"num_tokens": 35496076.0,
|
| 282 |
+
"step": 1250
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"entropy": 1.477535102367401,
|
| 286 |
+
"epoch": 0.4421768707482993,
|
| 287 |
+
"grad_norm": 1.140625,
|
| 288 |
+
"learning_rate": 1.9863510173539193e-05,
|
| 289 |
+
"loss": 1.1849075317382813,
|
| 290 |
+
"mean_token_accuracy": 0.6999362814426422,
|
| 291 |
+
"num_tokens": 36924227.0,
|
| 292 |
+
"step": 1300
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"entropy": 1.4699360299110413,
|
| 296 |
+
"epoch": 0.45918367346938777,
|
| 297 |
+
"grad_norm": 0.9921875,
|
| 298 |
+
"learning_rate": 1.982900310021936e-05,
|
| 299 |
+
"loss": 1.1758278656005858,
|
| 300 |
+
"mean_token_accuracy": 0.7016613483428955,
|
| 301 |
+
"num_tokens": 38298695.0,
|
| 302 |
+
"step": 1350
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"entropy": 1.4796976780891418,
|
| 306 |
+
"epoch": 0.47619047619047616,
|
| 307 |
+
"grad_norm": 1.2421875,
|
| 308 |
+
"learning_rate": 1.9790648306958504e-05,
|
| 309 |
+
"loss": 1.1764393615722657,
|
| 310 |
+
"mean_token_accuracy": 0.7006231868267059,
|
| 311 |
+
"num_tokens": 39730904.0,
|
| 312 |
+
"step": 1400
|
| 313 |
+
},
|
| 314 |
+
{
|
| 315 |
+
"entropy": 1.4659403300285339,
|
| 316 |
+
"epoch": 0.4931972789115646,
|
| 317 |
+
"grad_norm": 1.109375,
|
| 318 |
+
"learning_rate": 1.974846080835184e-05,
|
| 319 |
+
"loss": 1.1739279174804687,
|
| 320 |
+
"mean_token_accuracy": 0.7042246878147125,
|
| 321 |
+
"num_tokens": 41159605.0,
|
| 322 |
+
"step": 1450
|
| 323 |
+
},
|
| 324 |
+
{
|
| 325 |
+
"entropy": 1.4660640573501587,
|
| 326 |
+
"epoch": 0.5102040816326531,
|
| 327 |
+
"grad_norm": 1.15625,
|
| 328 |
+
"learning_rate": 1.9702457119368175e-05,
|
| 329 |
+
"loss": 1.1742501831054688,
|
| 330 |
+
"mean_token_accuracy": 0.7028728783130646,
|
| 331 |
+
"num_tokens": 42591337.0,
|
| 332 |
+
"step": 1500
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 0.5102040816326531,
|
| 336 |
+
"eval_entropy": 1.4500642824573677,
|
| 337 |
+
"eval_loss": 1.1341153383255005,
|
| 338 |
+
"eval_mean_token_accuracy": 0.7110618072397569,
|
| 339 |
+
"eval_num_tokens": 42591337.0,
|
| 340 |
+
"eval_runtime": 9.205,
|
| 341 |
+
"eval_samples_per_second": 102.988,
|
| 342 |
+
"eval_steps_per_second": 12.928,
|
| 343 |
+
"step": 1500
|
| 344 |
+
},
|
| 345 |
+
{
|
| 346 |
+
"entropy": 1.4817289400100708,
|
| 347 |
+
"epoch": 0.5272108843537415,
|
| 348 |
+
"grad_norm": 1.15625,
|
| 349 |
+
"learning_rate": 1.965265524888488e-05,
|
| 350 |
+
"loss": 1.1780020904541015,
|
| 351 |
+
"mean_token_accuracy": 0.7016154682636261,
|
| 352 |
+
"num_tokens": 44005399.0,
|
| 353 |
+
"step": 1550
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"entropy": 1.4686012029647828,
|
| 357 |
+
"epoch": 0.54421768707483,
|
| 358 |
+
"grad_norm": 1.0859375,
|
| 359 |
+
"learning_rate": 1.9599074692638007e-05,
|
| 360 |
+
"loss": 1.1757928466796874,
|
| 361 |
+
"mean_token_accuracy": 0.7019401288032532,
|
| 362 |
+
"num_tokens": 45411127.0,
|
| 363 |
+
"step": 1600
|
| 364 |
+
},
|
| 365 |
+
{
|
| 366 |
+
"entropy": 1.4470409440994263,
|
| 367 |
+
"epoch": 0.5612244897959183,
|
| 368 |
+
"grad_norm": 1.234375,
|
| 369 |
+
"learning_rate": 1.9541736425590398e-05,
|
| 370 |
+
"loss": 1.1487213897705078,
|
| 371 |
+
"mean_token_accuracy": 0.708456654548645,
|
| 372 |
+
"num_tokens": 46828771.0,
|
| 373 |
+
"step": 1650
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"entropy": 1.4495110416412353,
|
| 377 |
+
"epoch": 0.5782312925170068,
|
| 378 |
+
"grad_norm": 1.1484375,
|
| 379 |
+
"learning_rate": 1.9480662893720672e-05,
|
| 380 |
+
"loss": 1.1448258972167968,
|
| 381 |
+
"mean_token_accuracy": 0.7081741154193878,
|
| 382 |
+
"num_tokens": 48222827.0,
|
| 383 |
+
"step": 1700
|
| 384 |
+
},
|
| 385 |
+
{
|
| 386 |
+
"entropy": 1.4605646800994874,
|
| 387 |
+
"epoch": 0.5952380952380952,
|
| 388 |
+
"grad_norm": 1.125,
|
| 389 |
+
"learning_rate": 1.9415878005236413e-05,
|
| 390 |
+
"loss": 1.1873452758789063,
|
| 391 |
+
"mean_token_accuracy": 0.6996394228935242,
|
| 392 |
+
"num_tokens": 49670122.0,
|
| 393 |
+
"step": 1750
|
| 394 |
+
},
|
| 395 |
+
{
|
| 396 |
+
"entropy": 1.4478923654556275,
|
| 397 |
+
"epoch": 0.6122448979591837,
|
| 398 |
+
"grad_norm": 1.1484375,
|
| 399 |
+
"learning_rate": 1.9347407121214917e-05,
|
| 400 |
+
"loss": 1.1648704528808593,
|
| 401 |
+
"mean_token_accuracy": 0.7041294479370117,
|
| 402 |
+
"num_tokens": 51095588.0,
|
| 403 |
+
"step": 1800
|
| 404 |
+
},
|
| 405 |
+
{
|
| 406 |
+
"entropy": 1.4468252277374267,
|
| 407 |
+
"epoch": 0.6292517006802721,
|
| 408 |
+
"grad_norm": 1.0703125,
|
| 409 |
+
"learning_rate": 1.927527704567517e-05,
|
| 410 |
+
"loss": 1.1468747711181642,
|
| 411 |
+
"mean_token_accuracy": 0.7075002765655518,
|
| 412 |
+
"num_tokens": 52520201.0,
|
| 413 |
+
"step": 1850
|
| 414 |
+
},
|
| 415 |
+
{
|
| 416 |
+
"entropy": 1.4675183987617493,
|
| 417 |
+
"epoch": 0.6462585034013606,
|
| 418 |
+
"grad_norm": 1.078125,
|
| 419 |
+
"learning_rate": 1.9199516015085025e-05,
|
| 420 |
+
"loss": 1.162491912841797,
|
| 421 |
+
"mean_token_accuracy": 0.7048231554031372,
|
| 422 |
+
"num_tokens": 53944548.0,
|
| 423 |
+
"step": 1900
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"entropy": 1.4320320653915406,
|
| 427 |
+
"epoch": 0.6632653061224489,
|
| 428 |
+
"grad_norm": 1.1640625,
|
| 429 |
+
"learning_rate": 1.912015368730756e-05,
|
| 430 |
+
"loss": 1.1547863006591796,
|
| 431 |
+
"mean_token_accuracy": 0.7064202272891998,
|
| 432 |
+
"num_tokens": 55352104.0,
|
| 433 |
+
"step": 1950
|
| 434 |
+
},
|
| 435 |
+
{
|
| 436 |
+
"entropy": 1.4646633243560792,
|
| 437 |
+
"epoch": 0.6802721088435374,
|
| 438 |
+
"grad_norm": 1.0625,
|
| 439 |
+
"learning_rate": 1.9037221129991045e-05,
|
| 440 |
+
"loss": 1.1570500183105468,
|
| 441 |
+
"mean_token_accuracy": 0.7064848113059997,
|
| 442 |
+
"num_tokens": 56766925.0,
|
| 443 |
+
"step": 2000
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.6802721088435374,
|
| 447 |
+
"eval_entropy": 1.4422022985810994,
|
| 448 |
+
"eval_loss": 1.1238017082214355,
|
| 449 |
+
"eval_mean_token_accuracy": 0.7135469041952566,
|
| 450 |
+
"eval_num_tokens": 56766925.0,
|
| 451 |
+
"eval_runtime": 8.3815,
|
| 452 |
+
"eval_samples_per_second": 113.106,
|
| 453 |
+
"eval_steps_per_second": 14.198,
|
| 454 |
+
"step": 2000
|
| 455 |
+
},
|
| 456 |
+
{
|
| 457 |
+
"entropy": 1.4543561935424805,
|
| 458 |
+
"epoch": 0.6972789115646258,
|
| 459 |
+
"grad_norm": 1.21875,
|
| 460 |
+
"learning_rate": 1.8950750808407016e-05,
|
| 461 |
+
"loss": 1.1518563079833983,
|
| 462 |
+
"mean_token_accuracy": 0.7068261563777923,
|
| 463 |
+
"num_tokens": 58185378.0,
|
| 464 |
+
"step": 2050
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"entropy": 1.4502357959747314,
|
| 468 |
+
"epoch": 0.7142857142857143,
|
| 469 |
+
"grad_norm": 1.0,
|
| 470 |
+
"learning_rate": 1.8860776572741214e-05,
|
| 471 |
+
"loss": 1.1643141174316407,
|
| 472 |
+
"mean_token_accuracy": 0.7048910784721375,
|
| 473 |
+
"num_tokens": 59574639.0,
|
| 474 |
+
"step": 2100
|
| 475 |
+
},
|
| 476 |
+
{
|
| 477 |
+
"entropy": 1.4497376680374146,
|
| 478 |
+
"epoch": 0.7312925170068028,
|
| 479 |
+
"grad_norm": 1.1171875,
|
| 480 |
+
"learning_rate": 1.8767333644842414e-05,
|
| 481 |
+
"loss": 1.1560958862304687,
|
| 482 |
+
"mean_token_accuracy": 0.705675152540207,
|
| 483 |
+
"num_tokens": 60979229.0,
|
| 484 |
+
"step": 2150
|
| 485 |
+
},
|
| 486 |
+
{
|
| 487 |
+
"entropy": 1.4455422043800354,
|
| 488 |
+
"epoch": 0.7482993197278912,
|
| 489 |
+
"grad_norm": 1.1015625,
|
| 490 |
+
"learning_rate": 1.8670458604434253e-05,
|
| 491 |
+
"loss": 1.1813956451416017,
|
| 492 |
+
"mean_token_accuracy": 0.6994379079341888,
|
| 493 |
+
"num_tokens": 62398557.0,
|
| 494 |
+
"step": 2200
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"entropy": 1.4504392910003663,
|
| 498 |
+
"epoch": 0.7653061224489796,
|
| 499 |
+
"grad_norm": 1.2734375,
|
| 500 |
+
"learning_rate": 1.857018937479551e-05,
|
| 501 |
+
"loss": 1.163807601928711,
|
| 502 |
+
"mean_token_accuracy": 0.7050833916664123,
|
| 503 |
+
"num_tokens": 63824693.0,
|
| 504 |
+
"step": 2250
|
| 505 |
+
},
|
| 506 |
+
{
|
| 507 |
+
"entropy": 1.4483773112297058,
|
| 508 |
+
"epoch": 0.782312925170068,
|
| 509 |
+
"grad_norm": 1.03125,
|
| 510 |
+
"learning_rate": 1.846656520791446e-05,
|
| 511 |
+
"loss": 1.1639373779296875,
|
| 512 |
+
"mean_token_accuracy": 0.7042666506767273,
|
| 513 |
+
"num_tokens": 65262483.0,
|
| 514 |
+
"step": 2300
|
| 515 |
+
},
|
| 516 |
+
{
|
| 517 |
+
"entropy": 1.4623826432228089,
|
| 518 |
+
"epoch": 0.7993197278911565,
|
| 519 |
+
"grad_norm": 1.1171875,
|
| 520 |
+
"learning_rate": 1.835962666912303e-05,
|
| 521 |
+
"loss": 1.1614888763427735,
|
| 522 |
+
"mean_token_accuracy": 0.7056748032569885,
|
| 523 |
+
"num_tokens": 66699392.0,
|
| 524 |
+
"step": 2350
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"entropy": 1.4484982657432557,
|
| 528 |
+
"epoch": 0.8163265306122449,
|
| 529 |
+
"grad_norm": 0.94921875,
|
| 530 |
+
"learning_rate": 1.8249415621216853e-05,
|
| 531 |
+
"loss": 1.164166717529297,
|
| 532 |
+
"mean_token_accuracy": 0.7041208696365356,
|
| 533 |
+
"num_tokens": 68163337.0,
|
| 534 |
+
"step": 2400
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"entropy": 1.4403035855293274,
|
| 538 |
+
"epoch": 0.8333333333333334,
|
| 539 |
+
"grad_norm": 1.171875,
|
| 540 |
+
"learning_rate": 1.8135975208067442e-05,
|
| 541 |
+
"loss": 1.1633216857910156,
|
| 542 |
+
"mean_token_accuracy": 0.7039952325820923,
|
| 543 |
+
"num_tokens": 69584260.0,
|
| 544 |
+
"step": 2450
|
| 545 |
+
},
|
| 546 |
+
{
|
| 547 |
+
"entropy": 1.4662260293960572,
|
| 548 |
+
"epoch": 0.8503401360544217,
|
| 549 |
+
"grad_norm": 1.1484375,
|
| 550 |
+
"learning_rate": 1.80193498377328e-05,
|
| 551 |
+
"loss": 1.1636495208740234,
|
| 552 |
+
"mean_token_accuracy": 0.7047837376594543,
|
| 553 |
+
"num_tokens": 70984379.0,
|
| 554 |
+
"step": 2500
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 0.8503401360544217,
|
| 558 |
+
"eval_entropy": 1.4377584517502986,
|
| 559 |
+
"eval_loss": 1.1168923377990723,
|
| 560 |
+
"eval_mean_token_accuracy": 0.7147301846191663,
|
| 561 |
+
"eval_num_tokens": 70984379.0,
|
| 562 |
+
"eval_runtime": 7.3849,
|
| 563 |
+
"eval_samples_per_second": 128.37,
|
| 564 |
+
"eval_steps_per_second": 16.114,
|
| 565 |
+
"step": 2500
|
| 566 |
+
},
|
| 567 |
+
{
|
| 568 |
+
"entropy": 1.443549563884735,
|
| 569 |
+
"epoch": 0.8673469387755102,
|
| 570 |
+
"grad_norm": 1.046875,
|
| 571 |
+
"learning_rate": 1.7899585165073182e-05,
|
| 572 |
+
"loss": 1.148828353881836,
|
| 573 |
+
"mean_token_accuracy": 0.708294941186905,
|
| 574 |
+
"num_tokens": 72388922.0,
|
| 575 |
+
"step": 2550
|
| 576 |
+
},
|
| 577 |
+
{
|
| 578 |
+
"entropy": 1.461836290359497,
|
| 579 |
+
"epoch": 0.8843537414965986,
|
| 580 |
+
"grad_norm": 1.015625,
|
| 581 |
+
"learning_rate": 1.77767280738788e-05,
|
| 582 |
+
"loss": 1.18120361328125,
|
| 583 |
+
"mean_token_accuracy": 0.7006327843666077,
|
| 584 |
+
"num_tokens": 73820763.0,
|
| 585 |
+
"step": 2600
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"entropy": 1.4670860075950622,
|
| 589 |
+
"epoch": 0.9013605442176871,
|
| 590 |
+
"grad_norm": 1.1015625,
|
| 591 |
+
"learning_rate": 1.7650826658516375e-05,
|
| 592 |
+
"loss": 1.190813751220703,
|
| 593 |
+
"mean_token_accuracy": 0.6987618601322174,
|
| 594 |
+
"num_tokens": 75254805.0,
|
| 595 |
+
"step": 2650
|
| 596 |
+
},
|
| 597 |
+
{
|
| 598 |
+
"entropy": 1.4515372467041017,
|
| 599 |
+
"epoch": 0.9183673469387755,
|
| 600 |
+
"grad_norm": 0.9140625,
|
| 601 |
+
"learning_rate": 1.7521930205101857e-05,
|
| 602 |
+
"loss": 1.170417251586914,
|
| 603 |
+
"mean_token_accuracy": 0.7032169091701508,
|
| 604 |
+
"num_tokens": 76693263.0,
|
| 605 |
+
"step": 2700
|
| 606 |
+
},
|
| 607 |
+
{
|
| 608 |
+
"entropy": 1.4385710954666138,
|
| 609 |
+
"epoch": 0.935374149659864,
|
| 610 |
+
"grad_norm": 1.0390625,
|
| 611 |
+
"learning_rate": 1.7390089172206594e-05,
|
| 612 |
+
"loss": 1.1422464752197266,
|
| 613 |
+
"mean_token_accuracy": 0.7097269940376282,
|
| 614 |
+
"num_tokens": 78106463.0,
|
| 615 |
+
"step": 2750
|
| 616 |
+
},
|
| 617 |
+
{
|
| 618 |
+
"entropy": 1.4470832967758178,
|
| 619 |
+
"epoch": 0.9523809523809523,
|
| 620 |
+
"grad_norm": 1.09375,
|
| 621 |
+
"learning_rate": 1.7255355171104492e-05,
|
| 622 |
+
"loss": 1.169587173461914,
|
| 623 |
+
"mean_token_accuracy": 0.7027925884723664,
|
| 624 |
+
"num_tokens": 79532081.0,
|
| 625 |
+
"step": 2800
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"entropy": 1.462675015926361,
|
| 629 |
+
"epoch": 0.9693877551020408,
|
| 630 |
+
"grad_norm": 1.03125,
|
| 631 |
+
"learning_rate": 1.711778094556799e-05,
|
| 632 |
+
"loss": 1.1886711883544923,
|
| 633 |
+
"mean_token_accuracy": 0.6994748151302338,
|
| 634 |
+
"num_tokens": 80999149.0,
|
| 635 |
+
"step": 2850
|
| 636 |
+
},
|
| 637 |
+
{
|
| 638 |
+
"entropy": 1.4251791214942933,
|
| 639 |
+
"epoch": 0.9863945578231292,
|
| 640 |
+
"grad_norm": 0.953125,
|
| 641 |
+
"learning_rate": 1.6977420351220656e-05,
|
| 642 |
+
"loss": 1.138572998046875,
|
| 643 |
+
"mean_token_accuracy": 0.7102291858196259,
|
| 644 |
+
"num_tokens": 82436061.0,
|
| 645 |
+
"step": 2900
|
| 646 |
+
},
|
| 647 |
+
{
|
| 648 |
+
"entropy": 1.4440323853492736,
|
| 649 |
+
"epoch": 1.0034013605442176,
|
| 650 |
+
"grad_norm": 1.171875,
|
| 651 |
+
"learning_rate": 1.683432833445453e-05,
|
| 652 |
+
"loss": 1.1582012176513672,
|
| 653 |
+
"mean_token_accuracy": 0.7026605701446533,
|
| 654 |
+
"num_tokens": 83831706.0,
|
| 655 |
+
"step": 2950
|
| 656 |
+
},
|
| 657 |
+
{
|
| 658 |
+
"entropy": 1.4328801035881042,
|
| 659 |
+
"epoch": 1.0204081632653061,
|
| 660 |
+
"grad_norm": 1.046875,
|
| 661 |
+
"learning_rate": 1.668856091092051e-05,
|
| 662 |
+
"loss": 1.1449398040771483,
|
| 663 |
+
"mean_token_accuracy": 0.7084557473659515,
|
| 664 |
+
"num_tokens": 85243788.0,
|
| 665 |
+
"step": 3000
|
| 666 |
+
},
|
| 667 |
+
{
|
| 668 |
+
"epoch": 1.0204081632653061,
|
| 669 |
+
"eval_entropy": 1.425075779442026,
|
| 670 |
+
"eval_loss": 1.112339973449707,
|
| 671 |
+
"eval_mean_token_accuracy": 0.7156877617876068,
|
| 672 |
+
"eval_num_tokens": 85243788.0,
|
| 673 |
+
"eval_runtime": 9.148,
|
| 674 |
+
"eval_samples_per_second": 103.629,
|
| 675 |
+
"eval_steps_per_second": 13.008,
|
| 676 |
+
"step": 3000
|
| 677 |
+
},
|
| 678 |
+
{
|
| 679 |
+
"entropy": 1.457193055152893,
|
| 680 |
+
"epoch": 1.0374149659863945,
|
| 681 |
+
"grad_norm": 1.0078125,
|
| 682 |
+
"learning_rate": 1.6540175143600124e-05,
|
| 683 |
+
"loss": 1.1573681640625,
|
| 684 |
+
"mean_token_accuracy": 0.706193675994873,
|
| 685 |
+
"num_tokens": 86631509.0,
|
| 686 |
+
"step": 3050
|
| 687 |
+
},
|
| 688 |
+
{
|
| 689 |
+
"entropy": 1.4502310729026795,
|
| 690 |
+
"epoch": 1.054421768707483,
|
| 691 |
+
"grad_norm": 1.125,
|
| 692 |
+
"learning_rate": 1.6389229120467304e-05,
|
| 693 |
+
"loss": 1.1431642913818358,
|
| 694 |
+
"mean_token_accuracy": 0.7098660218715668,
|
| 695 |
+
"num_tokens": 88045606.0,
|
| 696 |
+
"step": 3100
|
| 697 |
+
},
|
| 698 |
+
{
|
| 699 |
+
"entropy": 1.4366277551651,
|
| 700 |
+
"epoch": 1.0714285714285714,
|
| 701 |
+
"grad_norm": 0.9453125,
|
| 702 |
+
"learning_rate": 1.6235781931748923e-05,
|
| 703 |
+
"loss": 1.1489845275878907,
|
| 704 |
+
"mean_token_accuracy": 0.707208307981491,
|
| 705 |
+
"num_tokens": 89467532.0,
|
| 706 |
+
"step": 3150
|
| 707 |
+
},
|
| 708 |
+
{
|
| 709 |
+
"entropy": 1.4366546082496643,
|
| 710 |
+
"epoch": 1.08843537414966,
|
| 711 |
+
"grad_norm": 1.1875,
|
| 712 |
+
"learning_rate": 1.6079893646793016e-05,
|
| 713 |
+
"loss": 1.1411451721191406,
|
| 714 |
+
"mean_token_accuracy": 0.7094867730140686,
|
| 715 |
+
"num_tokens": 90884635.0,
|
| 716 |
+
"step": 3200
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"entropy": 1.451173803806305,
|
| 720 |
+
"epoch": 1.1054421768707483,
|
| 721 |
+
"grad_norm": 0.98828125,
|
| 722 |
+
"learning_rate": 1.5921625290553628e-05,
|
| 723 |
+
"loss": 1.168065185546875,
|
| 724 |
+
"mean_token_accuracy": 0.7023441231250763,
|
| 725 |
+
"num_tokens": 92294540.0,
|
| 726 |
+
"step": 3250
|
| 727 |
+
},
|
| 728 |
+
{
|
| 729 |
+
"entropy": 1.4395431089401245,
|
| 730 |
+
"epoch": 1.1224489795918366,
|
| 731 |
+
"grad_norm": 1.0703125,
|
| 732 |
+
"learning_rate": 1.5761038819701658e-05,
|
| 733 |
+
"loss": 1.1530582427978515,
|
| 734 |
+
"mean_token_accuracy": 0.7062457311153412,
|
| 735 |
+
"num_tokens": 93722473.0,
|
| 736 |
+
"step": 3300
|
| 737 |
+
},
|
| 738 |
+
{
|
| 739 |
+
"entropy": 1.4319015264511108,
|
| 740 |
+
"epoch": 1.1394557823129252,
|
| 741 |
+
"grad_norm": 1.078125,
|
| 742 |
+
"learning_rate": 1.5598197098370928e-05,
|
| 743 |
+
"loss": 1.135885009765625,
|
| 744 |
+
"mean_token_accuracy": 0.7098275601863862,
|
| 745 |
+
"num_tokens": 95130619.0,
|
| 746 |
+
"step": 3350
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"entropy": 1.4478510427474975,
|
| 750 |
+
"epoch": 1.1564625850340136,
|
| 751 |
+
"grad_norm": 1.125,
|
| 752 |
+
"learning_rate": 1.5433163873548994e-05,
|
| 753 |
+
"loss": 1.1470938110351563,
|
| 754 |
+
"mean_token_accuracy": 0.7073779857158661,
|
| 755 |
+
"num_tokens": 96554760.0,
|
| 756 |
+
"step": 3400
|
| 757 |
+
},
|
| 758 |
+
{
|
| 759 |
+
"entropy": 1.4374464821815491,
|
| 760 |
+
"epoch": 1.1734693877551021,
|
| 761 |
+
"grad_norm": 1.1328125,
|
| 762 |
+
"learning_rate": 1.5266003750122394e-05,
|
| 763 |
+
"loss": 1.1405145263671874,
|
| 764 |
+
"mean_token_accuracy": 0.7093812191486358,
|
| 765 |
+
"num_tokens": 97955334.0,
|
| 766 |
+
"step": 3450
|
| 767 |
+
},
|
| 768 |
+
{
|
| 769 |
+
"entropy": 1.4236741828918458,
|
| 770 |
+
"epoch": 1.1904761904761905,
|
| 771 |
+
"grad_norm": 1.046875,
|
| 772 |
+
"learning_rate": 1.5096782165586037e-05,
|
| 773 |
+
"loss": 1.1330892181396484,
|
| 774 |
+
"mean_token_accuracy": 0.7110612964630127,
|
| 775 |
+
"num_tokens": 99395276.0,
|
| 776 |
+
"step": 3500
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 1.1904761904761905,
|
| 780 |
+
"eval_entropy": 1.4220578530255485,
|
| 781 |
+
"eval_loss": 1.1096009016036987,
|
| 782 |
+
"eval_mean_token_accuracy": 0.7162127835409982,
|
| 783 |
+
"eval_num_tokens": 99395276.0,
|
| 784 |
+
"eval_runtime": 7.5607,
|
| 785 |
+
"eval_samples_per_second": 125.385,
|
| 786 |
+
"eval_steps_per_second": 15.739,
|
| 787 |
+
"step": 3500
|
| 788 |
+
},
|
| 789 |
+
{
|
| 790 |
+
"entropy": 1.4502542996406556,
|
| 791 |
+
"epoch": 1.2074829931972788,
|
| 792 |
+
"grad_norm": 1.109375,
|
| 793 |
+
"learning_rate": 1.4925565364426626e-05,
|
| 794 |
+
"loss": 1.1476077270507812,
|
| 795 |
+
"mean_token_accuracy": 0.7080324268341065,
|
| 796 |
+
"num_tokens": 100822779.0,
|
| 797 |
+
"step": 3550
|
| 798 |
+
},
|
| 799 |
+
{
|
| 800 |
+
"entropy": 1.4335950636863708,
|
| 801 |
+
"epoch": 1.2244897959183674,
|
| 802 |
+
"grad_norm": 1.015625,
|
| 803 |
+
"learning_rate": 1.4752420372190235e-05,
|
| 804 |
+
"loss": 1.1536802673339843,
|
| 805 |
+
"mean_token_accuracy": 0.7067959046363831,
|
| 806 |
+
"num_tokens": 102255057.0,
|
| 807 |
+
"step": 3600
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"entropy": 1.431325912475586,
|
| 811 |
+
"epoch": 1.2414965986394557,
|
| 812 |
+
"grad_norm": 1.1171875,
|
| 813 |
+
"learning_rate": 1.4577414969244052e-05,
|
| 814 |
+
"loss": 1.1357704162597657,
|
| 815 |
+
"mean_token_accuracy": 0.7090205514431,
|
| 816 |
+
"num_tokens": 103682059.0,
|
| 817 |
+
"step": 3650
|
| 818 |
+
},
|
| 819 |
+
{
|
| 820 |
+
"entropy": 1.430570056438446,
|
| 821 |
+
"epoch": 1.2585034013605443,
|
| 822 |
+
"grad_norm": 1.0625,
|
| 823 |
+
"learning_rate": 1.4400617664242706e-05,
|
| 824 |
+
"loss": 1.1330258178710937,
|
| 825 |
+
"mean_token_accuracy": 0.7101845383644104,
|
| 826 |
+
"num_tokens": 105094949.0,
|
| 827 |
+
"step": 3700
|
| 828 |
+
},
|
| 829 |
+
{
|
| 830 |
+
"entropy": 1.4420754599571228,
|
| 831 |
+
"epoch": 1.2755102040816326,
|
| 832 |
+
"grad_norm": 0.96875,
|
| 833 |
+
"learning_rate": 1.422209766730945e-05,
|
| 834 |
+
"loss": 1.1578455352783203,
|
| 835 |
+
"mean_token_accuracy": 0.7062867367267609,
|
| 836 |
+
"num_tokens": 106525361.0,
|
| 837 |
+
"step": 3750
|
| 838 |
+
},
|
| 839 |
+
{
|
| 840 |
+
"entropy": 1.4218295288085938,
|
| 841 |
+
"epoch": 1.2925170068027212,
|
| 842 |
+
"grad_norm": 0.8984375,
|
| 843 |
+
"learning_rate": 1.4041924862942744e-05,
|
| 844 |
+
"loss": 1.1394439697265626,
|
| 845 |
+
"mean_token_accuracy": 0.7095408570766449,
|
| 846 |
+
"num_tokens": 107938288.0,
|
| 847 |
+
"step": 3800
|
| 848 |
+
},
|
| 849 |
+
{
|
| 850 |
+
"entropy": 1.425144808292389,
|
| 851 |
+
"epoch": 1.3095238095238095,
|
| 852 |
+
"grad_norm": 1.1875,
|
| 853 |
+
"learning_rate": 1.3860169782658863e-05,
|
| 854 |
+
"loss": 1.1289046478271485,
|
| 855 |
+
"mean_token_accuracy": 0.7117006468772888,
|
| 856 |
+
"num_tokens": 109343324.0,
|
| 857 |
+
"step": 3850
|
| 858 |
+
},
|
| 859 |
+
{
|
| 860 |
+
"entropy": 1.4213409280776979,
|
| 861 |
+
"epoch": 1.3265306122448979,
|
| 862 |
+
"grad_norm": 1.0625,
|
| 863 |
+
"learning_rate": 1.3676903577381231e-05,
|
| 864 |
+
"loss": 1.124376220703125,
|
| 865 |
+
"mean_token_accuracy": 0.7127799916267396,
|
| 866 |
+
"num_tokens": 110773342.0,
|
| 867 |
+
"step": 3900
|
| 868 |
+
},
|
| 869 |
+
{
|
| 870 |
+
"entropy": 1.430138614177704,
|
| 871 |
+
"epoch": 1.3435374149659864,
|
| 872 |
+
"grad_norm": 1.03125,
|
| 873 |
+
"learning_rate": 1.3492197989587221e-05,
|
| 874 |
+
"loss": 1.1303221893310547,
|
| 875 |
+
"mean_token_accuracy": 0.7113751316070557,
|
| 876 |
+
"num_tokens": 112223937.0,
|
| 877 |
+
"step": 3950
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"entropy": 1.421439278125763,
|
| 881 |
+
"epoch": 1.3605442176870748,
|
| 882 |
+
"grad_norm": 1.0390625,
|
| 883 |
+
"learning_rate": 1.3306125325223433e-05,
|
| 884 |
+
"loss": 1.128666305541992,
|
| 885 |
+
"mean_token_accuracy": 0.711964373588562,
|
| 886 |
+
"num_tokens": 113638897.0,
|
| 887 |
+
"step": 4000
|
| 888 |
+
},
|
| 889 |
+
{
|
| 890 |
+
"epoch": 1.3605442176870748,
|
| 891 |
+
"eval_entropy": 1.4151852511558212,
|
| 892 |
+
"eval_loss": 1.107796311378479,
|
| 893 |
+
"eval_mean_token_accuracy": 0.7166504459220822,
|
| 894 |
+
"eval_num_tokens": 113638897.0,
|
| 895 |
+
"eval_runtime": 8.7246,
|
| 896 |
+
"eval_samples_per_second": 108.658,
|
| 897 |
+
"eval_steps_per_second": 13.64,
|
| 898 |
+
"step": 4000
|
| 899 |
+
},
|
| 900 |
+
{
|
| 901 |
+
"entropy": 1.422073357105255,
|
| 902 |
+
"epoch": 1.3775510204081631,
|
| 903 |
+
"grad_norm": 1.1015625,
|
| 904 |
+
"learning_rate": 1.31187584254004e-05,
|
| 905 |
+
"loss": 1.137821807861328,
|
| 906 |
+
"mean_token_accuracy": 0.7102219009399414,
|
| 907 |
+
"num_tokens": 115030025.0,
|
| 908 |
+
"step": 4050
|
| 909 |
+
},
|
| 910 |
+
{
|
| 911 |
+
"entropy": 1.4335832858085633,
|
| 912 |
+
"epoch": 1.3945578231292517,
|
| 913 |
+
"grad_norm": 1.1171875,
|
| 914 |
+
"learning_rate": 1.2930170637877741e-05,
|
| 915 |
+
"loss": 1.1326773834228516,
|
| 916 |
+
"mean_token_accuracy": 0.7117250227928161,
|
| 917 |
+
"num_tokens": 116464440.0,
|
| 918 |
+
"step": 4100
|
| 919 |
+
},
|
| 920 |
+
{
|
| 921 |
+
"entropy": 1.432119870185852,
|
| 922 |
+
"epoch": 1.4115646258503403,
|
| 923 |
+
"grad_norm": 1.0546875,
|
| 924 |
+
"learning_rate": 1.2740435788351031e-05,
|
| 925 |
+
"loss": 1.144314727783203,
|
| 926 |
+
"mean_token_accuracy": 0.7074739003181457,
|
| 927 |
+
"num_tokens": 117895570.0,
|
| 928 |
+
"step": 4150
|
| 929 |
+
},
|
| 930 |
+
{
|
| 931 |
+
"entropy": 1.4382991909980773,
|
| 932 |
+
"epoch": 1.4285714285714286,
|
| 933 |
+
"grad_norm": 1.0546875,
|
| 934 |
+
"learning_rate": 1.254962815155155e-05,
|
| 935 |
+
"loss": 1.1475949859619141,
|
| 936 |
+
"mean_token_accuracy": 0.7079149842262268,
|
| 937 |
+
"num_tokens": 119315574.0,
|
| 938 |
+
"step": 4200
|
| 939 |
+
},
|
| 940 |
+
{
|
| 941 |
+
"entropy": 1.3948691391944885,
|
| 942 |
+
"epoch": 1.445578231292517,
|
| 943 |
+
"grad_norm": 0.99609375,
|
| 944 |
+
"learning_rate": 1.2357822422170266e-05,
|
| 945 |
+
"loss": 1.1130772399902344,
|
| 946 |
+
"mean_token_accuracy": 0.7150168895721436,
|
| 947 |
+
"num_tokens": 120717868.0,
|
| 948 |
+
"step": 4250
|
| 949 |
+
},
|
| 950 |
+
{
|
| 951 |
+
"entropy": 1.4280868649482727,
|
| 952 |
+
"epoch": 1.4625850340136055,
|
| 953 |
+
"grad_norm": 1.015625,
|
| 954 |
+
"learning_rate": 1.2165093685617377e-05,
|
| 955 |
+
"loss": 1.1200074005126952,
|
| 956 |
+
"mean_token_accuracy": 0.7135682916641235,
|
| 957 |
+
"num_tokens": 122146924.0,
|
| 958 |
+
"step": 4300
|
| 959 |
+
},
|
| 960 |
+
{
|
| 961 |
+
"entropy": 1.4316187024116516,
|
| 962 |
+
"epoch": 1.4795918367346939,
|
| 963 |
+
"grad_norm": 1.0390625,
|
| 964 |
+
"learning_rate": 1.1971517388628972e-05,
|
| 965 |
+
"loss": 1.1323360443115233,
|
| 966 |
+
"mean_token_accuracy": 0.7125037395954132,
|
| 967 |
+
"num_tokens": 123574594.0,
|
| 968 |
+
"step": 4350
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"entropy": 1.4233829760551453,
|
| 972 |
+
"epoch": 1.4965986394557822,
|
| 973 |
+
"grad_norm": 1.0625,
|
| 974 |
+
"learning_rate": 1.1777169309732178e-05,
|
| 975 |
+
"loss": 1.144235610961914,
|
| 976 |
+
"mean_token_accuracy": 0.707960147857666,
|
| 977 |
+
"num_tokens": 124985723.0,
|
| 978 |
+
"step": 4400
|
| 979 |
+
},
|
| 980 |
+
{
|
| 981 |
+
"entropy": 1.446813395023346,
|
| 982 |
+
"epoch": 1.5136054421768708,
|
| 983 |
+
"grad_norm": 0.9765625,
|
| 984 |
+
"learning_rate": 1.1582125529580437e-05,
|
| 985 |
+
"loss": 1.1620960235595703,
|
| 986 |
+
"mean_token_accuracy": 0.7044985663890838,
|
| 987 |
+
"num_tokens": 126444415.0,
|
| 988 |
+
"step": 4450
|
| 989 |
+
},
|
| 990 |
+
{
|
| 991 |
+
"entropy": 1.4292146420478822,
|
| 992 |
+
"epoch": 1.5306122448979593,
|
| 993 |
+
"grad_norm": 1.1328125,
|
| 994 |
+
"learning_rate": 1.1386462401170562e-05,
|
| 995 |
+
"loss": 1.1338434600830078,
|
| 996 |
+
"mean_token_accuracy": 0.7091768038272858,
|
| 997 |
+
"num_tokens": 127870485.0,
|
| 998 |
+
"step": 4500
|
| 999 |
+
},
|
| 1000 |
+
{
|
| 1001 |
+
"epoch": 1.5306122448979593,
|
| 1002 |
+
"eval_entropy": 1.4135069837089347,
|
| 1003 |
+
"eval_loss": 1.106415033340454,
|
| 1004 |
+
"eval_mean_token_accuracy": 0.7169664757592338,
|
| 1005 |
+
"eval_num_tokens": 127870485.0,
|
| 1006 |
+
"eval_runtime": 8.6098,
|
| 1007 |
+
"eval_samples_per_second": 110.107,
|
| 1008 |
+
"eval_steps_per_second": 13.821,
|
| 1009 |
+
"step": 4500
|
| 1010 |
+
},
|
| 1011 |
+
{
|
| 1012 |
+
"entropy": 1.412301104068756,
|
| 1013 |
+
"epoch": 1.5476190476190477,
|
| 1014 |
+
"grad_norm": 1.1171875,
|
| 1015 |
+
"learning_rate": 1.1190256519953097e-05,
|
| 1016 |
+
"loss": 1.1280068206787108,
|
| 1017 |
+
"mean_token_accuracy": 0.7108085489273072,
|
| 1018 |
+
"num_tokens": 129318094.0,
|
| 1019 |
+
"step": 4550
|
| 1020 |
+
},
|
| 1021 |
+
{
|
| 1022 |
+
"entropy": 1.4223233222961427,
|
| 1023 |
+
"epoch": 1.564625850340136,
|
| 1024 |
+
"grad_norm": 1.1484375,
|
| 1025 |
+
"learning_rate": 1.0993584693847827e-05,
|
| 1026 |
+
"loss": 1.1158720397949218,
|
| 1027 |
+
"mean_token_accuracy": 0.7139767467975616,
|
| 1028 |
+
"num_tokens": 130726607.0,
|
| 1029 |
+
"step": 4600
|
| 1030 |
+
},
|
| 1031 |
+
{
|
| 1032 |
+
"entropy": 1.42511239528656,
|
| 1033 |
+
"epoch": 1.5816326530612246,
|
| 1034 |
+
"grad_norm": 1.1171875,
|
| 1035 |
+
"learning_rate": 1.0796523913176092e-05,
|
| 1036 |
+
"loss": 1.1147970581054687,
|
| 1037 |
+
"mean_token_accuracy": 0.7152622473239899,
|
| 1038 |
+
"num_tokens": 132145387.0,
|
| 1039 |
+
"step": 4650
|
| 1040 |
+
},
|
| 1041 |
+
{
|
| 1042 |
+
"entropy": 1.4350685811042785,
|
| 1043 |
+
"epoch": 1.598639455782313,
|
| 1044 |
+
"grad_norm": 1.0625,
|
| 1045 |
+
"learning_rate": 1.0599151320521703e-05,
|
| 1046 |
+
"loss": 1.1316419982910155,
|
| 1047 |
+
"mean_token_accuracy": 0.7108878707885742,
|
| 1048 |
+
"num_tokens": 133560830.0,
|
| 1049 |
+
"step": 4700
|
| 1050 |
+
},
|
| 1051 |
+
{
|
| 1052 |
+
"entropy": 1.4368190813064574,
|
| 1053 |
+
"epoch": 1.6156462585034013,
|
| 1054 |
+
"grad_norm": 1.0703125,
|
| 1055 |
+
"learning_rate": 1.040154418053223e-05,
|
| 1056 |
+
"loss": 1.1656169128417968,
|
| 1057 |
+
"mean_token_accuracy": 0.7034963488578796,
|
| 1058 |
+
"num_tokens": 134992182.0,
|
| 1059 |
+
"step": 4750
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"entropy": 1.4219933152198792,
|
| 1063 |
+
"epoch": 1.6326530612244898,
|
| 1064 |
+
"grad_norm": 1.0234375,
|
| 1065 |
+
"learning_rate": 1.0203779849672546e-05,
|
| 1066 |
+
"loss": 1.1332920837402343,
|
| 1067 |
+
"mean_token_accuracy": 0.7100002634525299,
|
| 1068 |
+
"num_tokens": 136426899.0,
|
| 1069 |
+
"step": 4800
|
| 1070 |
+
},
|
| 1071 |
+
{
|
| 1072 |
+
"entropy": 1.4121305012702943,
|
| 1073 |
+
"epoch": 1.6496598639455784,
|
| 1074 |
+
"grad_norm": 1.1484375,
|
| 1075 |
+
"learning_rate": 1.0005935745942395e-05,
|
| 1076 |
+
"loss": 1.120484619140625,
|
| 1077 |
+
"mean_token_accuracy": 0.7137669050693511,
|
| 1078 |
+
"num_tokens": 137821124.0,
|
| 1079 |
+
"step": 4850
|
| 1080 |
+
},
|
| 1081 |
+
{
|
| 1082 |
+
"entropy": 1.4223187255859375,
|
| 1083 |
+
"epoch": 1.6666666666666665,
|
| 1084 |
+
"grad_norm": 1.21875,
|
| 1085 |
+
"learning_rate": 9.808089318569872e-06,
|
| 1086 |
+
"loss": 1.13340087890625,
|
| 1087 |
+
"mean_token_accuracy": 0.7105760216712952,
|
| 1088 |
+
"num_tokens": 139238831.0,
|
| 1089 |
+
"step": 4900
|
| 1090 |
+
},
|
| 1091 |
+
{
|
| 1092 |
+
"entropy": 1.4339079999923705,
|
| 1093 |
+
"epoch": 1.683673469387755,
|
| 1094 |
+
"grad_norm": 1.1171875,
|
| 1095 |
+
"learning_rate": 9.610318017692717e-06,
|
| 1096 |
+
"loss": 1.131591567993164,
|
| 1097 |
+
"mean_token_accuracy": 0.7117097544670105,
|
| 1098 |
+
"num_tokens": 140670083.0,
|
| 1099 |
+
"step": 4950
|
| 1100 |
+
},
|
| 1101 |
+
{
|
| 1102 |
+
"entropy": 1.42487220287323,
|
| 1103 |
+
"epoch": 1.7006802721088436,
|
| 1104 |
+
"grad_norm": 1.0,
|
| 1105 |
+
"learning_rate": 9.412699264039187e-06,
|
| 1106 |
+
"loss": 1.1361915588378906,
|
| 1107 |
+
"mean_token_accuracy": 0.709000414609909,
|
| 1108 |
+
"num_tokens": 142109640.0,
|
| 1109 |
+
"step": 5000
|
| 1110 |
+
},
|
| 1111 |
+
{
|
| 1112 |
+
"epoch": 1.7006802721088436,
|
| 1113 |
+
"eval_entropy": 1.4130272725049187,
|
| 1114 |
+
"eval_loss": 1.105644941329956,
|
| 1115 |
+
"eval_mean_token_accuracy": 0.7170687387971317,
|
| 1116 |
+
"eval_num_tokens": 142109640.0,
|
| 1117 |
+
"eval_runtime": 9.1754,
|
| 1118 |
+
"eval_samples_per_second": 103.32,
|
| 1119 |
+
"eval_steps_per_second": 12.97,
|
| 1120 |
+
"step": 5000
|
| 1121 |
+
},
|
| 1122 |
+
{
|
| 1123 |
+
"entropy": 1.43812607049942,
|
| 1124 |
+
"epoch": 1.717687074829932,
|
| 1125 |
+
"grad_norm": 1.109375,
|
| 1126 |
+
"learning_rate": 9.215310418620513e-06,
|
| 1127 |
+
"loss": 1.1505773162841797,
|
| 1128 |
+
"mean_token_accuracy": 0.7071945500373841,
|
| 1129 |
+
"num_tokens": 143529976.0,
|
| 1130 |
+
"step": 5050
|
| 1131 |
+
},
|
| 1132 |
+
{
|
| 1133 |
+
"entropy": 1.4412685680389403,
|
| 1134 |
+
"epoch": 1.7346938775510203,
|
| 1135 |
+
"grad_norm": 1.2890625,
|
| 1136 |
+
"learning_rate": 9.018228752446671e-06,
|
| 1137 |
+
"loss": 1.1481983947753907,
|
| 1138 |
+
"mean_token_accuracy": 0.7087244749069214,
|
| 1139 |
+
"num_tokens": 144937854.0,
|
| 1140 |
+
"step": 5100
|
| 1141 |
+
},
|
| 1142 |
+
{
|
| 1143 |
+
"entropy": 1.4139077353477478,
|
| 1144 |
+
"epoch": 1.751700680272109,
|
| 1145 |
+
"grad_norm": 1.046875,
|
| 1146 |
+
"learning_rate": 8.821531416277427e-06,
|
| 1147 |
+
"loss": 1.1202762603759766,
|
| 1148 |
+
"mean_token_accuracy": 0.7141082406044006,
|
| 1149 |
+
"num_tokens": 146358885.0,
|
| 1150 |
+
"step": 5150
|
| 1151 |
+
},
|
| 1152 |
+
{
|
| 1153 |
+
"entropy": 1.4265229201316834,
|
| 1154 |
+
"epoch": 1.7687074829931972,
|
| 1155 |
+
"grad_norm": 1.0390625,
|
| 1156 |
+
"learning_rate": 8.625295410420451e-06,
|
| 1157 |
+
"loss": 1.1334566497802734,
|
| 1158 |
+
"mean_token_accuracy": 0.7119065308570862,
|
| 1159 |
+
"num_tokens": 147797164.0,
|
| 1160 |
+
"step": 5200
|
| 1161 |
+
},
|
| 1162 |
+
{
|
| 1163 |
+
"entropy": 1.4280506682395935,
|
| 1164 |
+
"epoch": 1.7857142857142856,
|
| 1165 |
+
"grad_norm": 1.125,
|
| 1166 |
+
"learning_rate": 8.429597554588291e-06,
|
| 1167 |
+
"loss": 1.1369068908691407,
|
| 1168 |
+
"mean_token_accuracy": 0.7097280561923981,
|
| 1169 |
+
"num_tokens": 149201547.0,
|
| 1170 |
+
"step": 5250
|
| 1171 |
+
},
|
| 1172 |
+
{
|
| 1173 |
+
"entropy": 1.4155354285240174,
|
| 1174 |
+
"epoch": 1.8027210884353742,
|
| 1175 |
+
"grad_norm": 1.0390625,
|
| 1176 |
+
"learning_rate": 8.234514457826101e-06,
|
| 1177 |
+
"loss": 1.1342711639404297,
|
| 1178 |
+
"mean_token_accuracy": 0.7102336215972901,
|
| 1179 |
+
"num_tokens": 150625894.0,
|
| 1180 |
+
"step": 5300
|
| 1181 |
+
},
|
| 1182 |
+
{
|
| 1183 |
+
"entropy": 1.4110008382797241,
|
| 1184 |
+
"epoch": 1.8197278911564627,
|
| 1185 |
+
"grad_norm": 1.1640625,
|
| 1186 |
+
"learning_rate": 8.040122488521802e-06,
|
| 1187 |
+
"loss": 1.1284471893310546,
|
| 1188 |
+
"mean_token_accuracy": 0.7116459333896636,
|
| 1189 |
+
"num_tokens": 152068477.0,
|
| 1190 |
+
"step": 5350
|
| 1191 |
+
},
|
| 1192 |
+
{
|
| 1193 |
+
"entropy": 1.432766444683075,
|
| 1194 |
+
"epoch": 1.836734693877551,
|
| 1195 |
+
"grad_norm": 1.109375,
|
| 1196 |
+
"learning_rate": 7.84649774451044e-06,
|
| 1197 |
+
"loss": 1.1442298126220702,
|
| 1198 |
+
"mean_token_accuracy": 0.7084469604492187,
|
| 1199 |
+
"num_tokens": 153452805.0,
|
| 1200 |
+
"step": 5400
|
| 1201 |
+
},
|
| 1202 |
+
{
|
| 1203 |
+
"entropy": 1.4228289747238159,
|
| 1204 |
+
"epoch": 1.8537414965986394,
|
| 1205 |
+
"grad_norm": 1.0,
|
| 1206 |
+
"learning_rate": 7.653716023284498e-06,
|
| 1207 |
+
"loss": 1.1224955749511718,
|
| 1208 |
+
"mean_token_accuracy": 0.7124292123317718,
|
| 1209 |
+
"num_tokens": 154873137.0,
|
| 1210 |
+
"step": 5450
|
| 1211 |
+
},
|
| 1212 |
+
{
|
| 1213 |
+
"entropy": 1.4096328234672546,
|
| 1214 |
+
"epoch": 1.870748299319728,
|
| 1215 |
+
"grad_norm": 1.109375,
|
| 1216 |
+
"learning_rate": 7.46185279232176e-06,
|
| 1217 |
+
"loss": 1.1197357940673829,
|
| 1218 |
+
"mean_token_accuracy": 0.7131345391273498,
|
| 1219 |
+
"num_tokens": 156324700.0,
|
| 1220 |
+
"step": 5500
|
| 1221 |
+
},
|
| 1222 |
+
{
|
| 1223 |
+
"epoch": 1.870748299319728,
|
| 1224 |
+
"eval_entropy": 1.4128466764418017,
|
| 1225 |
+
"eval_loss": 1.1052268743515015,
|
| 1226 |
+
"eval_mean_token_accuracy": 0.7171510763528968,
|
| 1227 |
+
"eval_num_tokens": 156324700.0,
|
| 1228 |
+
"eval_runtime": 9.0481,
|
| 1229 |
+
"eval_samples_per_second": 104.773,
|
| 1230 |
+
"eval_steps_per_second": 13.152,
|
| 1231 |
+
"step": 5500
|
| 1232 |
+
},
|
| 1233 |
+
{
|
| 1234 |
+
"entropy": 1.4428221464157105,
|
| 1235 |
+
"epoch": 1.8877551020408163,
|
| 1236 |
+
"grad_norm": 0.96484375,
|
| 1237 |
+
"learning_rate": 7.27098315954234e-06,
|
| 1238 |
+
"loss": 1.1470510101318359,
|
| 1239 |
+
"mean_token_accuracy": 0.7073501777648926,
|
| 1240 |
+
"num_tokens": 157737046.0,
|
| 1241 |
+
"step": 5550
|
| 1242 |
+
},
|
| 1243 |
+
{
|
| 1244 |
+
"entropy": 1.4279747819900512,
|
| 1245 |
+
"epoch": 1.9047619047619047,
|
| 1246 |
+
"grad_norm": 1.0390625,
|
| 1247 |
+
"learning_rate": 7.081181843906513e-06,
|
| 1248 |
+
"loss": 1.1433404541015626,
|
| 1249 |
+
"mean_token_accuracy": 0.7082498788833618,
|
| 1250 |
+
"num_tokens": 159172992.0,
|
| 1251 |
+
"step": 5600
|
| 1252 |
+
},
|
| 1253 |
+
{
|
| 1254 |
+
"entropy": 1.4385703134536743,
|
| 1255 |
+
"epoch": 1.9217687074829932,
|
| 1256 |
+
"grad_norm": 1.1328125,
|
| 1257 |
+
"learning_rate": 6.892523146164782e-06,
|
| 1258 |
+
"loss": 1.1372301483154297,
|
| 1259 |
+
"mean_token_accuracy": 0.7099990272521972,
|
| 1260 |
+
"num_tokens": 160577822.0,
|
| 1261 |
+
"step": 5650
|
| 1262 |
+
},
|
| 1263 |
+
{
|
| 1264 |
+
"entropy": 1.4102177619934082,
|
| 1265 |
+
"epoch": 1.9387755102040818,
|
| 1266 |
+
"grad_norm": 1.0625,
|
| 1267 |
+
"learning_rate": 6.705080919771655e-06,
|
| 1268 |
+
"loss": 1.1329367065429687,
|
| 1269 |
+
"mean_token_accuracy": 0.7119775557518006,
|
| 1270 |
+
"num_tokens": 161991010.0,
|
| 1271 |
+
"step": 5700
|
| 1272 |
+
},
|
| 1273 |
+
{
|
| 1274 |
+
"entropy": 1.4429755330085754,
|
| 1275 |
+
"epoch": 1.95578231292517,
|
| 1276 |
+
"grad_norm": 1.125,
|
| 1277 |
+
"learning_rate": 6.51892854197454e-06,
|
| 1278 |
+
"loss": 1.1424610900878907,
|
| 1279 |
+
"mean_token_accuracy": 0.7083716082572937,
|
| 1280 |
+
"num_tokens": 163405551.0,
|
| 1281 |
+
"step": 5750
|
| 1282 |
+
},
|
| 1283 |
+
{
|
| 1284 |
+
"entropy": 1.4363023614883423,
|
| 1285 |
+
"epoch": 1.9727891156462585,
|
| 1286 |
+
"grad_norm": 1.109375,
|
| 1287 |
+
"learning_rate": 6.334138885089061e-06,
|
| 1288 |
+
"loss": 1.142376937866211,
|
| 1289 |
+
"mean_token_accuracy": 0.709513326883316,
|
| 1290 |
+
"num_tokens": 164828847.0,
|
| 1291 |
+
"step": 5800
|
| 1292 |
+
},
|
| 1293 |
+
{
|
| 1294 |
+
"entropy": 1.435585732460022,
|
| 1295 |
+
"epoch": 1.989795918367347,
|
| 1296 |
+
"grad_norm": 1.0234375,
|
| 1297 |
+
"learning_rate": 6.150784287972005e-06,
|
| 1298 |
+
"loss": 1.157130126953125,
|
| 1299 |
+
"mean_token_accuracy": 0.7060146403312683,
|
| 1300 |
+
"num_tokens": 166264927.0,
|
| 1301 |
+
"step": 5850
|
| 1302 |
+
},
|
| 1303 |
+
{
|
| 1304 |
+
"entropy": 1.4074757862091065,
|
| 1305 |
+
"epoch": 2.006802721088435,
|
| 1306 |
+
"grad_norm": 1.0546875,
|
| 1307 |
+
"learning_rate": 5.968936527703144e-06,
|
| 1308 |
+
"loss": 1.1172750091552734,
|
| 1309 |
+
"mean_token_accuracy": 0.7146967911720276,
|
| 1310 |
+
"num_tokens": 167673391.0,
|
| 1311 |
+
"step": 5900
|
| 1312 |
+
},
|
| 1313 |
+
{
|
| 1314 |
+
"entropy": 1.4369916796684266,
|
| 1315 |
+
"epoch": 2.0238095238095237,
|
| 1316 |
+
"grad_norm": 0.98828125,
|
| 1317 |
+
"learning_rate": 5.78866679148695e-06,
|
| 1318 |
+
"loss": 1.1662197875976563,
|
| 1319 |
+
"mean_token_accuracy": 0.7031553852558136,
|
| 1320 |
+
"num_tokens": 169094709.0,
|
| 1321 |
+
"step": 5950
|
| 1322 |
+
},
|
| 1323 |
+
{
|
| 1324 |
+
"entropy": 1.4330995082855225,
|
| 1325 |
+
"epoch": 2.0408163265306123,
|
| 1326 |
+
"grad_norm": 1.03125,
|
| 1327 |
+
"learning_rate": 5.610045648785225e-06,
|
| 1328 |
+
"loss": 1.1259635162353516,
|
| 1329 |
+
"mean_token_accuracy": 0.7129771506786347,
|
| 1330 |
+
"num_tokens": 170523291.0,
|
| 1331 |
+
"step": 6000
|
| 1332 |
+
},
|
| 1333 |
+
{
|
| 1334 |
+
"epoch": 2.0408163265306123,
|
| 1335 |
+
"eval_entropy": 1.4114271442429358,
|
| 1336 |
+
"eval_loss": 1.1050275564193726,
|
| 1337 |
+
"eval_mean_token_accuracy": 0.7171849367999229,
|
| 1338 |
+
"eval_num_tokens": 170523291.0,
|
| 1339 |
+
"eval_runtime": 9.0969,
|
| 1340 |
+
"eval_samples_per_second": 104.211,
|
| 1341 |
+
"eval_steps_per_second": 13.081,
|
| 1342 |
+
"step": 6000
|
| 1343 |
+
},
|
| 1344 |
+
{
|
| 1345 |
+
"entropy": 1.4260020923614503,
|
| 1346 |
+
"epoch": 2.057823129251701,
|
| 1347 |
+
"grad_norm": 1.078125,
|
| 1348 |
+
"learning_rate": 5.433143023691547e-06,
|
| 1349 |
+
"loss": 1.1178092193603515,
|
| 1350 |
+
"mean_token_accuracy": 0.7144936168193817,
|
| 1351 |
+
"num_tokens": 171938887.0,
|
| 1352 |
+
"step": 6050
|
| 1353 |
+
},
|
| 1354 |
+
{
|
| 1355 |
+
"entropy": 1.4517771482467652,
|
| 1356 |
+
"epoch": 2.074829931972789,
|
| 1357 |
+
"grad_norm": 1.0703125,
|
| 1358 |
+
"learning_rate": 5.2580281675584045e-06,
|
| 1359 |
+
"loss": 1.1548236083984376,
|
| 1360 |
+
"mean_token_accuracy": 0.7052713608741761,
|
| 1361 |
+
"num_tokens": 173371802.0,
|
| 1362 |
+
"step": 6100
|
| 1363 |
+
},
|
| 1364 |
+
{
|
| 1365 |
+
"entropy": 1.41314546585083,
|
| 1366 |
+
"epoch": 2.0918367346938775,
|
| 1367 |
+
"grad_norm": 1.0625,
|
| 1368 |
+
"learning_rate": 5.084769631887611e-06,
|
| 1369 |
+
"loss": 1.116751708984375,
|
| 1370 |
+
"mean_token_accuracy": 0.7143128430843353,
|
| 1371 |
+
"num_tokens": 174783437.0,
|
| 1372 |
+
"step": 6150
|
| 1373 |
+
},
|
| 1374 |
+
{
|
| 1375 |
+
"entropy": 1.4380582642555237,
|
| 1376 |
+
"epoch": 2.108843537414966,
|
| 1377 |
+
"grad_norm": 1.109375,
|
| 1378 |
+
"learning_rate": 4.913435241494749e-06,
|
| 1379 |
+
"loss": 1.14319091796875,
|
| 1380 |
+
"mean_token_accuracy": 0.7082942843437194,
|
| 1381 |
+
"num_tokens": 176235112.0,
|
| 1382 |
+
"step": 6200
|
| 1383 |
+
},
|
| 1384 |
+
{
|
| 1385 |
+
"entropy": 1.432828288078308,
|
| 1386 |
+
"epoch": 2.1258503401360542,
|
| 1387 |
+
"grad_norm": 1.046875,
|
| 1388 |
+
"learning_rate": 4.744092067958067e-06,
|
| 1389 |
+
"loss": 1.1341230010986327,
|
| 1390 |
+
"mean_token_accuracy": 0.7104381382465362,
|
| 1391 |
+
"num_tokens": 177657881.0,
|
| 1392 |
+
"step": 6250
|
| 1393 |
+
},
|
| 1394 |
+
{
|
| 1395 |
+
"entropy": 1.4256313157081604,
|
| 1396 |
+
"epoch": 2.142857142857143,
|
| 1397 |
+
"grad_norm": 1.0078125,
|
| 1398 |
+
"learning_rate": 4.5768064033622315e-06,
|
| 1399 |
+
"loss": 1.13789306640625,
|
| 1400 |
+
"mean_token_accuracy": 0.7089906525611878,
|
| 1401 |
+
"num_tokens": 179083511.0,
|
| 1402 |
+
"step": 6300
|
| 1403 |
+
},
|
| 1404 |
+
{
|
| 1405 |
+
"entropy": 1.4250177597999574,
|
| 1406 |
+
"epoch": 2.1598639455782314,
|
| 1407 |
+
"grad_norm": 1.2421875,
|
| 1408 |
+
"learning_rate": 4.411643734347228e-06,
|
| 1409 |
+
"loss": 1.142816925048828,
|
| 1410 |
+
"mean_token_accuracy": 0.7079160177707672,
|
| 1411 |
+
"num_tokens": 180486838.0,
|
| 1412 |
+
"step": 6350
|
| 1413 |
+
},
|
| 1414 |
+
{
|
| 1415 |
+
"entropy": 1.397556071281433,
|
| 1416 |
+
"epoch": 2.17687074829932,
|
| 1417 |
+
"grad_norm": 0.9921875,
|
| 1418 |
+
"learning_rate": 4.248668716472597e-06,
|
| 1419 |
+
"loss": 1.1062926483154296,
|
| 1420 |
+
"mean_token_accuracy": 0.716565626859665,
|
| 1421 |
+
"num_tokens": 181896305.0,
|
| 1422 |
+
"step": 6400
|
| 1423 |
+
},
|
| 1424 |
+
{
|
| 1425 |
+
"entropy": 1.4365149688720704,
|
| 1426 |
+
"epoch": 2.193877551020408,
|
| 1427 |
+
"grad_norm": 0.9609375,
|
| 1428 |
+
"learning_rate": 4.087945148906956e-06,
|
| 1429 |
+
"loss": 1.1448672485351563,
|
| 1430 |
+
"mean_token_accuracy": 0.7083494472503662,
|
| 1431 |
+
"num_tokens": 183287787.0,
|
| 1432 |
+
"step": 6450
|
| 1433 |
+
},
|
| 1434 |
+
{
|
| 1435 |
+
"entropy": 1.411766264438629,
|
| 1436 |
+
"epoch": 2.2108843537414966,
|
| 1437 |
+
"grad_norm": 1.21875,
|
| 1438 |
+
"learning_rate": 3.929535949452806e-06,
|
| 1439 |
+
"loss": 1.1034652709960937,
|
| 1440 |
+
"mean_token_accuracy": 0.7170097494125366,
|
| 1441 |
+
"num_tokens": 184697327.0,
|
| 1442 |
+
"step": 6500
|
| 1443 |
+
},
|
| 1444 |
+
{
|
| 1445 |
+
"epoch": 2.2108843537414966,
|
| 1446 |
+
"eval_entropy": 1.411220889131562,
|
| 1447 |
+
"eval_loss": 1.1048741340637207,
|
| 1448 |
+
"eval_mean_token_accuracy": 0.7172508204684538,
|
| 1449 |
+
"eval_num_tokens": 184697327.0,
|
| 1450 |
+
"eval_runtime": 8.3646,
|
| 1451 |
+
"eval_samples_per_second": 113.334,
|
| 1452 |
+
"eval_steps_per_second": 14.227,
|
| 1453 |
+
"step": 6500
|
| 1454 |
+
},
|
| 1455 |
+
{
|
| 1456 |
+
"entropy": 1.4167416214942932,
|
| 1457 |
+
"epoch": 2.227891156462585,
|
| 1458 |
+
"grad_norm": 1.0078125,
|
| 1459 |
+
"learning_rate": 3.773503129916374e-06,
|
| 1460 |
+
"loss": 1.1365108489990234,
|
| 1461 |
+
"mean_token_accuracy": 0.7105149948596954,
|
| 1462 |
+
"num_tokens": 186127393.0,
|
| 1463 |
+
"step": 6550
|
| 1464 |
+
},
|
| 1465 |
+
{
|
| 1466 |
+
"entropy": 1.4231563568115235,
|
| 1467 |
+
"epoch": 2.2448979591836733,
|
| 1468 |
+
"grad_norm": 1.046875,
|
| 1469 |
+
"learning_rate": 3.6199077718320817e-06,
|
| 1470 |
+
"loss": 1.146829605102539,
|
| 1471 |
+
"mean_token_accuracy": 0.7080222427845001,
|
| 1472 |
+
"num_tokens": 187593729.0,
|
| 1473 |
+
"step": 6600
|
| 1474 |
+
},
|
| 1475 |
+
{
|
| 1476 |
+
"entropy": 1.4201726150512695,
|
| 1477 |
+
"epoch": 2.261904761904762,
|
| 1478 |
+
"grad_norm": 1.046875,
|
| 1479 |
+
"learning_rate": 3.4688100025512052e-06,
|
| 1480 |
+
"loss": 1.1279085540771485,
|
| 1481 |
+
"mean_token_accuracy": 0.7107222032546997,
|
| 1482 |
+
"num_tokens": 188993851.0,
|
| 1483 |
+
"step": 6650
|
| 1484 |
+
},
|
| 1485 |
+
{
|
| 1486 |
+
"entropy": 1.414864149093628,
|
| 1487 |
+
"epoch": 2.2789115646258504,
|
| 1488 |
+
"grad_norm": 1.0234375,
|
| 1489 |
+
"learning_rate": 3.3202689717040805e-06,
|
| 1490 |
+
"loss": 1.1230371856689454,
|
| 1491 |
+
"mean_token_accuracy": 0.713964912891388,
|
| 1492 |
+
"num_tokens": 190427604.0,
|
| 1493 |
+
"step": 6700
|
| 1494 |
+
},
|
| 1495 |
+
{
|
| 1496 |
+
"entropy": 1.4270901560783387,
|
| 1497 |
+
"epoch": 2.295918367346939,
|
| 1498 |
+
"grad_norm": 1.078125,
|
| 1499 |
+
"learning_rate": 3.1743428280450217e-06,
|
| 1500 |
+
"loss": 1.1262576293945312,
|
| 1501 |
+
"mean_token_accuracy": 0.7121612048149109,
|
| 1502 |
+
"num_tokens": 191836070.0,
|
| 1503 |
+
"step": 6750
|
| 1504 |
+
},
|
| 1505 |
+
{
|
| 1506 |
+
"entropy": 1.4266995477676392,
|
| 1507 |
+
"epoch": 2.312925170068027,
|
| 1508 |
+
"grad_norm": 1.1953125,
|
| 1509 |
+
"learning_rate": 3.0310886966890695e-06,
|
| 1510 |
+
"loss": 1.1296147155761718,
|
| 1511 |
+
"mean_token_accuracy": 0.7117228031158447,
|
| 1512 |
+
"num_tokens": 193246580.0,
|
| 1513 |
+
"step": 6800
|
| 1514 |
+
},
|
| 1515 |
+
{
|
| 1516 |
+
"entropy": 1.4152509379386902,
|
| 1517 |
+
"epoch": 2.3299319727891157,
|
| 1518 |
+
"grad_norm": 1.078125,
|
| 1519 |
+
"learning_rate": 2.890562656749465e-06,
|
| 1520 |
+
"loss": 1.1302790069580078,
|
| 1521 |
+
"mean_token_accuracy": 0.7125261425971985,
|
| 1522 |
+
"num_tokens": 194684126.0,
|
| 1523 |
+
"step": 6850
|
| 1524 |
+
},
|
| 1525 |
+
{
|
| 1526 |
+
"entropy": 1.4296420979499818,
|
| 1527 |
+
"epoch": 2.3469387755102042,
|
| 1528 |
+
"grad_norm": 1.1171875,
|
| 1529 |
+
"learning_rate": 2.752819719384573e-06,
|
| 1530 |
+
"loss": 1.139165802001953,
|
| 1531 |
+
"mean_token_accuracy": 0.7097991979122162,
|
| 1532 |
+
"num_tokens": 196075134.0,
|
| 1533 |
+
"step": 6900
|
| 1534 |
+
},
|
| 1535 |
+
{
|
| 1536 |
+
"entropy": 1.4089247822761535,
|
| 1537 |
+
"epoch": 2.3639455782312924,
|
| 1538 |
+
"grad_norm": 1.0,
|
| 1539 |
+
"learning_rate": 2.6179138062628863e-06,
|
| 1540 |
+
"loss": 1.1174855041503906,
|
| 1541 |
+
"mean_token_accuracy": 0.7136220681667328,
|
| 1542 |
+
"num_tokens": 197502453.0,
|
| 1543 |
+
"step": 6950
|
| 1544 |
+
},
|
| 1545 |
+
{
|
| 1546 |
+
"entropy": 1.4326133465766906,
|
| 1547 |
+
"epoch": 2.380952380952381,
|
| 1548 |
+
"grad_norm": 1.078125,
|
| 1549 |
+
"learning_rate": 2.485897728454546e-06,
|
| 1550 |
+
"loss": 1.137528533935547,
|
| 1551 |
+
"mean_token_accuracy": 0.7098107814788819,
|
| 1552 |
+
"num_tokens": 198927675.0,
|
| 1553 |
+
"step": 7000
|
| 1554 |
+
},
|
| 1555 |
+
{
|
| 1556 |
+
"epoch": 2.380952380952381,
|
| 1557 |
+
"eval_entropy": 1.4109101956631958,
|
| 1558 |
+
"eval_loss": 1.1048496961593628,
|
| 1559 |
+
"eval_mean_token_accuracy": 0.7171741062853517,
|
| 1560 |
+
"eval_num_tokens": 198927675.0,
|
| 1561 |
+
"eval_runtime": 8.35,
|
| 1562 |
+
"eval_samples_per_second": 113.532,
|
| 1563 |
+
"eval_steps_per_second": 14.251,
|
| 1564 |
+
"step": 7000
|
| 1565 |
+
},
|
| 1566 |
+
{
|
| 1567 |
+
"entropy": 1.4286697244644164,
|
| 1568 |
+
"epoch": 2.3979591836734695,
|
| 1569 |
+
"grad_norm": 1.0390625,
|
| 1570 |
+
"learning_rate": 2.3568231657575834e-06,
|
| 1571 |
+
"loss": 1.129428939819336,
|
| 1572 |
+
"mean_token_accuracy": 0.7123888862133027,
|
| 1573 |
+
"num_tokens": 200323695.0,
|
| 1574 |
+
"step": 7050
|
| 1575 |
+
},
|
| 1576 |
+
{
|
| 1577 |
+
"entropy": 1.438240523338318,
|
| 1578 |
+
"epoch": 2.4149659863945576,
|
| 1579 |
+
"grad_norm": 0.94140625,
|
| 1580 |
+
"learning_rate": 2.2307406464670556e-06,
|
| 1581 |
+
"loss": 1.1379161071777344,
|
| 1582 |
+
"mean_token_accuracy": 0.7092709636688233,
|
| 1583 |
+
"num_tokens": 201726356.0,
|
| 1584 |
+
"step": 7100
|
| 1585 |
+
},
|
| 1586 |
+
{
|
| 1587 |
+
"entropy": 1.4395070052146912,
|
| 1588 |
+
"epoch": 2.431972789115646,
|
| 1589 |
+
"grad_norm": 1.0078125,
|
| 1590 |
+
"learning_rate": 2.1076995275949363e-06,
|
| 1591 |
+
"loss": 1.1427032470703125,
|
| 1592 |
+
"mean_token_accuracy": 0.7088804817199708,
|
| 1593 |
+
"num_tokens": 203117543.0,
|
| 1594 |
+
"step": 7150
|
| 1595 |
+
},
|
| 1596 |
+
{
|
| 1597 |
+
"entropy": 1.4223812985420228,
|
| 1598 |
+
"epoch": 2.4489795918367347,
|
| 1599 |
+
"grad_norm": 1.0859375,
|
| 1600 |
+
"learning_rate": 1.987747975548525e-06,
|
| 1601 |
+
"loss": 1.141911849975586,
|
| 1602 |
+
"mean_token_accuracy": 0.7085710716247559,
|
| 1603 |
+
"num_tokens": 204540538.0,
|
| 1604 |
+
"step": 7200
|
| 1605 |
+
},
|
| 1606 |
+
{
|
| 1607 |
+
"entropy": 1.4200545740127564,
|
| 1608 |
+
"epoch": 2.4659863945578233,
|
| 1609 |
+
"grad_norm": 1.1015625,
|
| 1610 |
+
"learning_rate": 1.8709329472749305e-06,
|
| 1611 |
+
"loss": 1.134008255004883,
|
| 1612 |
+
"mean_token_accuracy": 0.7102875554561615,
|
| 1613 |
+
"num_tokens": 205964981.0,
|
| 1614 |
+
"step": 7250
|
| 1615 |
+
},
|
| 1616 |
+
{
|
| 1617 |
+
"entropy": 1.4074384212493896,
|
| 1618 |
+
"epoch": 2.4829931972789114,
|
| 1619 |
+
"grad_norm": 1.0234375,
|
| 1620 |
+
"learning_rate": 1.757300171879045e-06,
|
| 1621 |
+
"loss": 1.1225076293945313,
|
| 1622 |
+
"mean_token_accuracy": 0.7127889108657837,
|
| 1623 |
+
"num_tokens": 207376354.0,
|
| 1624 |
+
"step": 7300
|
| 1625 |
+
},
|
| 1626 |
+
{
|
| 1627 |
+
"entropy": 1.4282367515563965,
|
| 1628 |
+
"epoch": 2.5,
|
| 1629 |
+
"grad_norm": 1.125,
|
| 1630 |
+
"learning_rate": 1.6468941327221377e-06,
|
| 1631 |
+
"loss": 1.1383217620849608,
|
| 1632 |
+
"mean_token_accuracy": 0.7087402617931366,
|
| 1633 |
+
"num_tokens": 208768190.0,
|
| 1634 |
+
"step": 7350
|
| 1635 |
+
},
|
| 1636 |
+
{
|
| 1637 |
+
"entropy": 1.4284853887557984,
|
| 1638 |
+
"epoch": 2.5170068027210886,
|
| 1639 |
+
"grad_norm": 0.98046875,
|
| 1640 |
+
"learning_rate": 1.539758050008141e-06,
|
| 1641 |
+
"loss": 1.1509806060791015,
|
| 1642 |
+
"mean_token_accuracy": 0.7068107867240906,
|
| 1643 |
+
"num_tokens": 210181751.0,
|
| 1644 |
+
"step": 7400
|
| 1645 |
+
},
|
| 1646 |
+
{
|
| 1647 |
+
"entropy": 1.441183557510376,
|
| 1648 |
+
"epoch": 2.534013605442177,
|
| 1649 |
+
"grad_norm": 1.0703125,
|
| 1650 |
+
"learning_rate": 1.4359338638644194e-06,
|
| 1651 |
+
"loss": 1.1412885284423828,
|
| 1652 |
+
"mean_token_accuracy": 0.7094214224815368,
|
| 1653 |
+
"num_tokens": 211592789.0,
|
| 1654 |
+
"step": 7450
|
| 1655 |
+
},
|
| 1656 |
+
{
|
| 1657 |
+
"entropy": 1.4259497380256654,
|
| 1658 |
+
"epoch": 2.5510204081632653,
|
| 1659 |
+
"grad_norm": 1.0703125,
|
| 1660 |
+
"learning_rate": 1.335462217923631e-06,
|
| 1661 |
+
"loss": 1.1345823669433595,
|
| 1662 |
+
"mean_token_accuracy": 0.7104575765132904,
|
| 1663 |
+
"num_tokens": 213028550.0,
|
| 1664 |
+
"step": 7500
|
| 1665 |
+
},
|
| 1666 |
+
{
|
| 1667 |
+
"epoch": 2.5510204081632653,
|
| 1668 |
+
"eval_entropy": 1.4108752523149763,
|
| 1669 |
+
"eval_loss": 1.1047948598861694,
|
| 1670 |
+
"eval_mean_token_accuracy": 0.7172804800402216,
|
| 1671 |
+
"eval_num_tokens": 213028550.0,
|
| 1672 |
+
"eval_runtime": 9.5013,
|
| 1673 |
+
"eval_samples_per_second": 99.775,
|
| 1674 |
+
"eval_steps_per_second": 12.525,
|
| 1675 |
+
"step": 7500
|
| 1676 |
+
},
|
| 1677 |
+
{
|
| 1678 |
+
"entropy": 1.4120974230766297,
|
| 1679 |
+
"epoch": 2.568027210884354,
|
| 1680 |
+
"grad_norm": 1.0625,
|
| 1681 |
+
"learning_rate": 1.2383824434131253e-06,
|
| 1682 |
+
"loss": 1.1174241638183593,
|
| 1683 |
+
"mean_token_accuracy": 0.7148010683059692,
|
| 1684 |
+
"num_tokens": 214461463.0,
|
| 1685 |
+
"step": 7550
|
| 1686 |
+
},
|
| 1687 |
+
{
|
| 1688 |
+
"entropy": 1.42849999666214,
|
| 1689 |
+
"epoch": 2.5850340136054424,
|
| 1690 |
+
"grad_norm": 1.0,
|
| 1691 |
+
"learning_rate": 1.1447325437581214e-06,
|
| 1692 |
+
"loss": 1.1466537475585938,
|
| 1693 |
+
"mean_token_accuracy": 0.7079911732673645,
|
| 1694 |
+
"num_tokens": 215914614.0,
|
| 1695 |
+
"step": 7600
|
| 1696 |
+
},
|
| 1697 |
+
{
|
| 1698 |
+
"entropy": 1.4341383004188537,
|
| 1699 |
+
"epoch": 2.6020408163265305,
|
| 1700 |
+
"grad_norm": 1.0390625,
|
| 1701 |
+
"learning_rate": 1.0545491797046438e-06,
|
| 1702 |
+
"loss": 1.1218763732910155,
|
| 1703 |
+
"mean_token_accuracy": 0.7130703330039978,
|
| 1704 |
+
"num_tokens": 217329434.0,
|
| 1705 |
+
"step": 7650
|
| 1706 |
+
},
|
| 1707 |
+
{
|
| 1708 |
+
"entropy": 1.4117230129241944,
|
| 1709 |
+
"epoch": 2.619047619047619,
|
| 1710 |
+
"grad_norm": 1.234375,
|
| 1711 |
+
"learning_rate": 9.678676549681077e-07,
|
| 1712 |
+
"loss": 1.1172777557373046,
|
| 1713 |
+
"mean_token_accuracy": 0.7140469861030578,
|
| 1714 |
+
"num_tokens": 218728299.0,
|
| 1715 |
+
"step": 7700
|
| 1716 |
+
},
|
| 1717 |
+
{
|
| 1718 |
+
"entropy": 1.4421565866470336,
|
| 1719 |
+
"epoch": 2.6360544217687076,
|
| 1720 |
+
"grad_norm": 1.1484375,
|
| 1721 |
+
"learning_rate": 8.84721902413097e-07,
|
| 1722 |
+
"loss": 1.1367300415039063,
|
| 1723 |
+
"mean_token_accuracy": 0.7104098796844482,
|
| 1724 |
+
"num_tokens": 220137703.0,
|
| 1725 |
+
"step": 7750
|
| 1726 |
+
},
|
| 1727 |
+
{
|
| 1728 |
+
"entropy": 1.4460611391067504,
|
| 1729 |
+
"epoch": 2.6530612244897958,
|
| 1730 |
+
"grad_norm": 0.97265625,
|
| 1731 |
+
"learning_rate": 8.05144470769823e-07,
|
| 1732 |
+
"loss": 1.1389141845703126,
|
| 1733 |
+
"mean_token_accuracy": 0.7093100786209107,
|
| 1734 |
+
"num_tokens": 221566274.0,
|
| 1735 |
+
"step": 7800
|
| 1736 |
+
},
|
| 1737 |
+
{
|
| 1738 |
+
"entropy": 1.433594093322754,
|
| 1739 |
+
"epoch": 2.6700680272108843,
|
| 1740 |
+
"grad_norm": 1.09375,
|
| 1741 |
+
"learning_rate": 7.291665118923919e-07,
|
| 1742 |
+
"loss": 1.1337957000732422,
|
| 1743 |
+
"mean_token_accuracy": 0.7100949645042419,
|
| 1744 |
+
"num_tokens": 223013879.0,
|
| 1745 |
+
"step": 7850
|
| 1746 |
+
},
|
| 1747 |
+
{
|
| 1748 |
+
"entropy": 1.4160945701599121,
|
| 1749 |
+
"epoch": 2.687074829931973,
|
| 1750 |
+
"grad_norm": 1.125,
|
| 1751 |
+
"learning_rate": 6.568177685639187e-07,
|
| 1752 |
+
"loss": 1.1077276611328124,
|
| 1753 |
+
"mean_token_accuracy": 0.7158815062046051,
|
| 1754 |
+
"num_tokens": 224430160.0,
|
| 1755 |
+
"step": 7900
|
| 1756 |
+
},
|
| 1757 |
+
{
|
| 1758 |
+
"entropy": 1.4393118023872375,
|
| 1759 |
+
"epoch": 2.704081632653061,
|
| 1760 |
+
"grad_norm": 1.03125,
|
| 1761 |
+
"learning_rate": 5.881265628532395e-07,
|
| 1762 |
+
"loss": 1.1686244201660156,
|
| 1763 |
+
"mean_token_accuracy": 0.700288165807724,
|
| 1764 |
+
"num_tokens": 225850566.0,
|
| 1765 |
+
"step": 7950
|
| 1766 |
+
},
|
| 1767 |
+
{
|
| 1768 |
+
"entropy": 1.418517336845398,
|
| 1769 |
+
"epoch": 2.7210884353741496,
|
| 1770 |
+
"grad_norm": 1.1171875,
|
| 1771 |
+
"learning_rate": 5.231197850277914e-07,
|
| 1772 |
+
"loss": 1.117198257446289,
|
| 1773 |
+
"mean_token_accuracy": 0.7138521051406861,
|
| 1774 |
+
"num_tokens": 227252331.0,
|
| 1775 |
+
"step": 8000
|
| 1776 |
+
},
|
| 1777 |
+
{
|
| 1778 |
+
"epoch": 2.7210884353741496,
|
| 1779 |
+
"eval_entropy": 1.411157799368145,
|
| 1780 |
+
"eval_loss": 1.1048526763916016,
|
| 1781 |
+
"eval_mean_token_accuracy": 0.7171973621144014,
|
| 1782 |
+
"eval_num_tokens": 227252331.0,
|
| 1783 |
+
"eval_runtime": 9.4712,
|
| 1784 |
+
"eval_samples_per_second": 100.093,
|
| 1785 |
+
"eval_steps_per_second": 12.564,
|
| 1786 |
+
"step": 8000
|
| 1787 |
+
},
|
| 1788 |
+
{
|
| 1789 |
+
"entropy": 1.4183010959625244,
|
| 1790 |
+
"epoch": 2.738095238095238,
|
| 1791 |
+
"grad_norm": 1.0078125,
|
| 1792 |
+
"learning_rate": 4.618228830269822e-07,
|
| 1793 |
+
"loss": 1.1152452850341796,
|
| 1794 |
+
"mean_token_accuracy": 0.7141678440570831,
|
| 1795 |
+
"num_tokens": 228659040.0,
|
| 1796 |
+
"step": 8050
|
| 1797 |
+
},
|
| 1798 |
+
{
|
| 1799 |
+
"entropy": 1.4264278268814088,
|
| 1800 |
+
"epoch": 2.7551020408163263,
|
| 1801 |
+
"grad_norm": 1.0,
|
| 1802 |
+
"learning_rate": 4.0425985250019174e-07,
|
| 1803 |
+
"loss": 1.141219711303711,
|
| 1804 |
+
"mean_token_accuracy": 0.7086251521110535,
|
| 1805 |
+
"num_tokens": 230087322.0,
|
| 1806 |
+
"step": 8100
|
| 1807 |
+
},
|
| 1808 |
+
{
|
| 1809 |
+
"entropy": 1.4321845197677612,
|
| 1810 |
+
"epoch": 2.772108843537415,
|
| 1811 |
+
"grad_norm": 1.078125,
|
| 1812 |
+
"learning_rate": 3.5045322741329435e-07,
|
| 1813 |
+
"loss": 1.1400668334960937,
|
| 1814 |
+
"mean_token_accuracy": 0.7088430786132812,
|
| 1815 |
+
"num_tokens": 231534171.0,
|
| 1816 |
+
"step": 8150
|
| 1817 |
+
},
|
| 1818 |
+
{
|
| 1819 |
+
"entropy": 1.438843650817871,
|
| 1820 |
+
"epoch": 2.7891156462585034,
|
| 1821 |
+
"grad_norm": 0.98046875,
|
| 1822 |
+
"learning_rate": 3.004240712273743e-07,
|
| 1823 |
+
"loss": 1.1484903717041015,
|
| 1824 |
+
"mean_token_accuracy": 0.7076469850540161,
|
| 1825 |
+
"num_tokens": 232975150.0,
|
| 1826 |
+
"step": 8200
|
| 1827 |
+
},
|
| 1828 |
+
{
|
| 1829 |
+
"entropy": 1.425174674987793,
|
| 1830 |
+
"epoch": 2.806122448979592,
|
| 1831 |
+
"grad_norm": 0.9140625,
|
| 1832 |
+
"learning_rate": 2.5419196865309916e-07,
|
| 1833 |
+
"loss": 1.1197071075439453,
|
| 1834 |
+
"mean_token_accuracy": 0.7135571861267089,
|
| 1835 |
+
"num_tokens": 234415037.0,
|
| 1836 |
+
"step": 8250
|
| 1837 |
+
},
|
| 1838 |
+
{
|
| 1839 |
+
"entropy": 1.414315857887268,
|
| 1840 |
+
"epoch": 2.8231292517006805,
|
| 1841 |
+
"grad_norm": 1.1796875,
|
| 1842 |
+
"learning_rate": 2.1177501798397947e-07,
|
| 1843 |
+
"loss": 1.1076421356201172,
|
| 1844 |
+
"mean_token_accuracy": 0.716143217086792,
|
| 1845 |
+
"num_tokens": 235839773.0,
|
| 1846 |
+
"step": 8300
|
| 1847 |
+
},
|
| 1848 |
+
{
|
| 1849 |
+
"entropy": 1.4245373177528382,
|
| 1850 |
+
"epoch": 2.8401360544217686,
|
| 1851 |
+
"grad_norm": 1.125,
|
| 1852 |
+
"learning_rate": 1.7318982401149265e-07,
|
| 1853 |
+
"loss": 1.126745147705078,
|
| 1854 |
+
"mean_token_accuracy": 0.7123495972156525,
|
| 1855 |
+
"num_tokens": 237240977.0,
|
| 1856 |
+
"step": 8350
|
| 1857 |
+
},
|
| 1858 |
+
{
|
| 1859 |
+
"entropy": 1.4268240427970886,
|
| 1860 |
+
"epoch": 2.857142857142857,
|
| 1861 |
+
"grad_norm": 0.9375,
|
| 1862 |
+
"learning_rate": 1.3845149152487803e-07,
|
| 1863 |
+
"loss": 1.1193339538574218,
|
| 1864 |
+
"mean_token_accuracy": 0.7143187963962555,
|
| 1865 |
+
"num_tokens": 238665599.0,
|
| 1866 |
+
"step": 8400
|
| 1867 |
+
},
|
| 1868 |
+
{
|
| 1869 |
+
"entropy": 1.4306807327270508,
|
| 1870 |
+
"epoch": 2.8741496598639458,
|
| 1871 |
+
"grad_norm": 1.140625,
|
| 1872 |
+
"learning_rate": 1.0757361939813137e-07,
|
| 1873 |
+
"loss": 1.12787353515625,
|
| 1874 |
+
"mean_token_accuracy": 0.7113662779331207,
|
| 1875 |
+
"num_tokens": 240099917.0,
|
| 1876 |
+
"step": 8450
|
| 1877 |
+
},
|
| 1878 |
+
{
|
| 1879 |
+
"entropy": 1.4395443654060365,
|
| 1880 |
+
"epoch": 2.891156462585034,
|
| 1881 |
+
"grad_norm": 1.171875,
|
| 1882 |
+
"learning_rate": 8.056829526649767e-08,
|
| 1883 |
+
"loss": 1.1395121002197266,
|
| 1884 |
+
"mean_token_accuracy": 0.7095162570476532,
|
| 1885 |
+
"num_tokens": 241504767.0,
|
| 1886 |
+
"step": 8500
|
| 1887 |
+
},
|
| 1888 |
+
{
|
| 1889 |
+
"epoch": 2.891156462585034,
|
| 1890 |
+
"eval_entropy": 1.4105532329623438,
|
| 1891 |
+
"eval_loss": 1.1048182249069214,
|
| 1892 |
+
"eval_mean_token_accuracy": 0.7172953571592059,
|
| 1893 |
+
"eval_num_tokens": 241504767.0,
|
| 1894 |
+
"eval_runtime": 9.5391,
|
| 1895 |
+
"eval_samples_per_second": 99.381,
|
| 1896 |
+
"eval_steps_per_second": 12.475,
|
| 1897 |
+
"step": 8500
|
| 1898 |
+
},
|
| 1899 |
+
{
|
| 1900 |
+
"entropy": 1.4125055813789367,
|
| 1901 |
+
"epoch": 2.9081632653061225,
|
| 1902 |
+
"grad_norm": 1.0078125,
|
| 1903 |
+
"learning_rate": 5.7446090794580636e-08,
|
| 1904 |
+
"loss": 1.1209735870361328,
|
| 1905 |
+
"mean_token_accuracy": 0.713252090215683,
|
| 1906 |
+
"num_tokens": 242946321.0,
|
| 1907 |
+
"step": 8550
|
| 1908 |
+
},
|
| 1909 |
+
{
|
| 1910 |
+
"entropy": 1.4286201095581055,
|
| 1911 |
+
"epoch": 2.925170068027211,
|
| 1912 |
+
"grad_norm": 1.0859375,
|
| 1913 |
+
"learning_rate": 3.821605753789648e-08,
|
| 1914 |
+
"loss": 1.1291080474853517,
|
| 1915 |
+
"mean_token_accuracy": 0.712302508354187,
|
| 1916 |
+
"num_tokens": 244378684.0,
|
| 1917 |
+
"step": 8600
|
| 1918 |
+
},
|
| 1919 |
+
{
|
| 1920 |
+
"entropy": 1.4226014256477355,
|
| 1921 |
+
"epoch": 2.942176870748299,
|
| 1922 |
+
"grad_norm": 1.015625,
|
| 1923 |
+
"learning_rate": 2.288572339949835e-08,
|
| 1924 |
+
"loss": 1.1374082946777344,
|
| 1925 |
+
"mean_token_accuracy": 0.7114299070835114,
|
| 1926 |
+
"num_tokens": 245806693.0,
|
| 1927 |
+
"step": 8650
|
| 1928 |
+
},
|
| 1929 |
+
{
|
| 1930 |
+
"entropy": 1.4223952269554139,
|
| 1931 |
+
"epoch": 2.9591836734693877,
|
| 1932 |
+
"grad_norm": 1.0859375,
|
| 1933 |
+
"learning_rate": 1.1461089683053595e-08,
|
| 1934 |
+
"loss": 1.1320287322998046,
|
| 1935 |
+
"mean_token_accuracy": 0.7105390858650208,
|
| 1936 |
+
"num_tokens": 247211598.0,
|
| 1937 |
+
"step": 8700
|
| 1938 |
+
},
|
| 1939 |
+
{
|
| 1940 |
+
"entropy": 1.4174312782287597,
|
| 1941 |
+
"epoch": 2.9761904761904763,
|
| 1942 |
+
"grad_norm": 1.0234375,
|
| 1943 |
+
"learning_rate": 3.946628743543013e-09,
|
| 1944 |
+
"loss": 1.1252450561523437,
|
| 1945 |
+
"mean_token_accuracy": 0.713299845457077,
|
| 1946 |
+
"num_tokens": 248623558.0,
|
| 1947 |
+
"step": 8750
|
| 1948 |
+
},
|
| 1949 |
+
{
|
| 1950 |
+
"entropy": 1.430122504234314,
|
| 1951 |
+
"epoch": 2.9931972789115644,
|
| 1952 |
+
"grad_norm": 1.1640625,
|
| 1953 |
+
"learning_rate": 3.452822364835573e-10,
|
| 1954 |
+
"loss": 1.142134246826172,
|
| 1955 |
+
"mean_token_accuracy": 0.7098831164836884,
|
| 1956 |
+
"num_tokens": 250064804.0,
|
| 1957 |
+
"step": 8800
|
| 1958 |
+
}
|
| 1959 |
+
],
|
| 1960 |
+
"logging_steps": 50,
|
| 1961 |
+
"max_steps": 8820,
|
| 1962 |
+
"num_input_tokens_seen": 0,
|
| 1963 |
+
"num_train_epochs": 3,
|
| 1964 |
+
"save_steps": 500,
|
| 1965 |
+
"stateful_callbacks": {
|
| 1966 |
+
"TrainerControl": {
|
| 1967 |
+
"args": {
|
| 1968 |
+
"should_epoch_stop": false,
|
| 1969 |
+
"should_evaluate": false,
|
| 1970 |
+
"should_log": false,
|
| 1971 |
+
"should_save": true,
|
| 1972 |
+
"should_training_stop": true
|
| 1973 |
+
},
|
| 1974 |
+
"attributes": {}
|
| 1975 |
+
}
|
| 1976 |
+
},
|
| 1977 |
+
"total_flos": 1.3775254383370368e+18,
|
| 1978 |
+
"train_batch_size": 32,
|
| 1979 |
+
"trial_name": null,
|
| 1980 |
+
"trial_params": null
|
| 1981 |
+
}
|
checkpoint-8820/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2309a8eab46e70d77a5e11e1a0025a30065cc758ddf39c81cb9137e50f03b60b
|
| 3 |
+
size 5649
|
eval_metrics.json
ADDED
|
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"eval_loss": 1.1048007011413574,
|
| 3 |
+
"eval_runtime": 8.5485,
|
| 4 |
+
"eval_samples_per_second": 110.897,
|
| 5 |
+
"eval_steps_per_second": 13.921
|
| 6 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"eval_loss": 1.1048007011413574,
|
| 3 |
+
"eval_runtime": 8.5485,
|
| 4 |
+
"eval_samples_per_second": 110.897,
|
| 5 |
+
"eval_steps_per_second": 13.921
|
| 6 |
+
}
|
final/config.json
ADDED
|
@@ -0,0 +1,32 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"LlamaForCausalLM"
|
| 4 |
+
],
|
| 5 |
+
"attention_bias": false,
|
| 6 |
+
"attention_dropout": 0.0,
|
| 7 |
+
"bos_token_id": 0,
|
| 8 |
+
"dtype": "bfloat16",
|
| 9 |
+
"eos_token_id": 0,
|
| 10 |
+
"head_dim": 64,
|
| 11 |
+
"hidden_act": "silu",
|
| 12 |
+
"hidden_size": 1280,
|
| 13 |
+
"initializer_range": 0.02,
|
| 14 |
+
"intermediate_size": 2816,
|
| 15 |
+
"max_position_embeddings": 2048,
|
| 16 |
+
"mlp_bias": false,
|
| 17 |
+
"model_type": "llama",
|
| 18 |
+
"num_attention_heads": 20,
|
| 19 |
+
"num_hidden_layers": 24,
|
| 20 |
+
"num_key_value_heads": 10,
|
| 21 |
+
"pad_token_id": 1,
|
| 22 |
+
"pretraining_tp": 1,
|
| 23 |
+
"rms_norm_eps": 1e-05,
|
| 24 |
+
"rope_parameters": {
|
| 25 |
+
"rope_theta": 10000.0,
|
| 26 |
+
"rope_type": "default"
|
| 27 |
+
},
|
| 28 |
+
"tie_word_embeddings": false,
|
| 29 |
+
"transformers_version": "5.3.0",
|
| 30 |
+
"use_cache": false,
|
| 31 |
+
"vocab_size": 50304
|
| 32 |
+
}
|
final/generation_config.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"bos_token_id": 0,
|
| 4 |
+
"eos_token_id": [
|
| 5 |
+
0
|
| 6 |
+
],
|
| 7 |
+
"pad_token_id": 1,
|
| 8 |
+
"transformers_version": "5.3.0"
|
| 9 |
+
}
|
final/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b74230cccba97ff7a7d718cc33463a9830d024351a5b2eb867a0285f2fbe8ea7
|
| 3 |
+
size 1012681552
|
final/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
final/tokenizer_config.json
ADDED
|
@@ -0,0 +1,21 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_prefix_space": false,
|
| 3 |
+
"backend": "tokenizers",
|
| 4 |
+
"bos_token": "<|endoftext|>",
|
| 5 |
+
"eos_token": "<|endoftext|>",
|
| 6 |
+
"errors": "replace",
|
| 7 |
+
"extra_special_tokens": [
|
| 8 |
+
"<|system|>",
|
| 9 |
+
"<|user|>",
|
| 10 |
+
"<|assistant|>",
|
| 11 |
+
"<|tool_call|>",
|
| 12 |
+
"<|tool_result|>",
|
| 13 |
+
"<|end|>"
|
| 14 |
+
],
|
| 15 |
+
"is_local": true,
|
| 16 |
+
"model_max_length": 1000000000000000019884624838656,
|
| 17 |
+
"pad_token": "<|padding|>",
|
| 18 |
+
"tokenizer_class": "GPTNeoXTokenizer",
|
| 19 |
+
"trim_offsets": true,
|
| 20 |
+
"unk_token": "<|endoftext|>"
|
| 21 |
+
}
|
final/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2309a8eab46e70d77a5e11e1a0025a30065cc758ddf39c81cb9137e50f03b60b
|
| 3 |
+
size 5649
|
train_metrics.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"train_runtime": 2757.2959,
|
| 3 |
+
"train_samples_per_second": 102.347,
|
| 4 |
+
"train_steps_per_second": 3.199,
|
| 5 |
+
"total_flos": 1.3775254383370368e+18,
|
| 6 |
+
"train_loss": 1.158577897164827,
|
| 7 |
+
"train_examples": 94067
|
| 8 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"total_flos": 1.3775254383370368e+18,
|
| 3 |
+
"train_examples": 94067,
|
| 4 |
+
"train_loss": 1.158577897164827,
|
| 5 |
+
"train_runtime": 2757.2959,
|
| 6 |
+
"train_samples_per_second": 102.347,
|
| 7 |
+
"train_steps_per_second": 3.199
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,1993 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": null,
|
| 3 |
+
"best_metric": null,
|
| 4 |
+
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 3.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 8820,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"entropy": 1.7698939561843872,
|
| 14 |
+
"epoch": 0.017006802721088437,
|
| 15 |
+
"grad_norm": 3.796875,
|
| 16 |
+
"learning_rate": 1.112372304199773e-06,
|
| 17 |
+
"loss": 1.5452156066894531,
|
| 18 |
+
"mean_token_accuracy": 0.6630511581897736,
|
| 19 |
+
"num_tokens": 1416855.0,
|
| 20 |
+
"step": 50
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"entropy": 1.765206663608551,
|
| 24 |
+
"epoch": 0.034013605442176874,
|
| 25 |
+
"grad_norm": 4.6875,
|
| 26 |
+
"learning_rate": 2.2474460839954596e-06,
|
| 27 |
+
"loss": 1.5233023071289062,
|
| 28 |
+
"mean_token_accuracy": 0.6625268197059632,
|
| 29 |
+
"num_tokens": 2861955.0,
|
| 30 |
+
"step": 100
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"entropy": 1.7640055298805237,
|
| 34 |
+
"epoch": 0.05102040816326531,
|
| 35 |
+
"grad_norm": 2.984375,
|
| 36 |
+
"learning_rate": 3.3825198637911465e-06,
|
| 37 |
+
"loss": 1.4435250854492188,
|
| 38 |
+
"mean_token_accuracy": 0.6652392935752869,
|
| 39 |
+
"num_tokens": 4279472.0,
|
| 40 |
+
"step": 150
|
| 41 |
+
},
|
| 42 |
+
{
|
| 43 |
+
"entropy": 1.7068938755989074,
|
| 44 |
+
"epoch": 0.06802721088435375,
|
| 45 |
+
"grad_norm": 1.3828125,
|
| 46 |
+
"learning_rate": 4.517593643586833e-06,
|
| 47 |
+
"loss": 1.3378631591796875,
|
| 48 |
+
"mean_token_accuracy": 0.673524329662323,
|
| 49 |
+
"num_tokens": 5707775.0,
|
| 50 |
+
"step": 200
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"entropy": 1.6458502221107483,
|
| 54 |
+
"epoch": 0.08503401360544217,
|
| 55 |
+
"grad_norm": 1.1015625,
|
| 56 |
+
"learning_rate": 5.652667423382521e-06,
|
| 57 |
+
"loss": 1.2989051818847657,
|
| 58 |
+
"mean_token_accuracy": 0.6784149742126465,
|
| 59 |
+
"num_tokens": 7118906.0,
|
| 60 |
+
"step": 250
|
| 61 |
+
},
|
| 62 |
+
{
|
| 63 |
+
"entropy": 1.6340991377830505,
|
| 64 |
+
"epoch": 0.10204081632653061,
|
| 65 |
+
"grad_norm": 1.21875,
|
| 66 |
+
"learning_rate": 6.787741203178207e-06,
|
| 67 |
+
"loss": 1.2801470947265625,
|
| 68 |
+
"mean_token_accuracy": 0.6825140416622162,
|
| 69 |
+
"num_tokens": 8529671.0,
|
| 70 |
+
"step": 300
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"entropy": 1.599980125427246,
|
| 74 |
+
"epoch": 0.11904761904761904,
|
| 75 |
+
"grad_norm": 1.1640625,
|
| 76 |
+
"learning_rate": 7.922814982973895e-06,
|
| 77 |
+
"loss": 1.2615231323242186,
|
| 78 |
+
"mean_token_accuracy": 0.6857551598548889,
|
| 79 |
+
"num_tokens": 9973401.0,
|
| 80 |
+
"step": 350
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"entropy": 1.570529818534851,
|
| 84 |
+
"epoch": 0.1360544217687075,
|
| 85 |
+
"grad_norm": 1.078125,
|
| 86 |
+
"learning_rate": 9.057888762769582e-06,
|
| 87 |
+
"loss": 1.2502473449707032,
|
| 88 |
+
"mean_token_accuracy": 0.6881939315795899,
|
| 89 |
+
"num_tokens": 11391611.0,
|
| 90 |
+
"step": 400
|
| 91 |
+
},
|
| 92 |
+
{
|
| 93 |
+
"entropy": 1.5739640021324157,
|
| 94 |
+
"epoch": 0.15306122448979592,
|
| 95 |
+
"grad_norm": 1.1640625,
|
| 96 |
+
"learning_rate": 1.0192962542565269e-05,
|
| 97 |
+
"loss": 1.2381304168701173,
|
| 98 |
+
"mean_token_accuracy": 0.6904172372817993,
|
| 99 |
+
"num_tokens": 12818742.0,
|
| 100 |
+
"step": 450
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"entropy": 1.542729480266571,
|
| 104 |
+
"epoch": 0.17006802721088435,
|
| 105 |
+
"grad_norm": 1.2421875,
|
| 106 |
+
"learning_rate": 1.1328036322360954e-05,
|
| 107 |
+
"loss": 1.2234064483642577,
|
| 108 |
+
"mean_token_accuracy": 0.6928502190113067,
|
| 109 |
+
"num_tokens": 14230694.0,
|
| 110 |
+
"step": 500
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.17006802721088435,
|
| 114 |
+
"eval_entropy": 1.5409097210699771,
|
| 115 |
+
"eval_loss": 1.1968441009521484,
|
| 116 |
+
"eval_mean_token_accuracy": 0.6989319805337602,
|
| 117 |
+
"eval_num_tokens": 14230694.0,
|
| 118 |
+
"eval_runtime": 9.3471,
|
| 119 |
+
"eval_samples_per_second": 101.422,
|
| 120 |
+
"eval_steps_per_second": 12.731,
|
| 121 |
+
"step": 500
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"entropy": 1.5507296323776245,
|
| 125 |
+
"epoch": 0.1870748299319728,
|
| 126 |
+
"grad_norm": 1.2265625,
|
| 127 |
+
"learning_rate": 1.246311010215664e-05,
|
| 128 |
+
"loss": 1.2286612701416015,
|
| 129 |
+
"mean_token_accuracy": 0.6924433386325837,
|
| 130 |
+
"num_tokens": 15656461.0,
|
| 131 |
+
"step": 550
|
| 132 |
+
},
|
| 133 |
+
{
|
| 134 |
+
"entropy": 1.5728167843818666,
|
| 135 |
+
"epoch": 0.20408163265306123,
|
| 136 |
+
"grad_norm": 1.21875,
|
| 137 |
+
"learning_rate": 1.3598183881952328e-05,
|
| 138 |
+
"loss": 1.2478004455566407,
|
| 139 |
+
"mean_token_accuracy": 0.6871815168857575,
|
| 140 |
+
"num_tokens": 17079797.0,
|
| 141 |
+
"step": 600
|
| 142 |
+
},
|
| 143 |
+
{
|
| 144 |
+
"entropy": 1.5617320823669434,
|
| 145 |
+
"epoch": 0.22108843537414966,
|
| 146 |
+
"grad_norm": 1.125,
|
| 147 |
+
"learning_rate": 1.4733257661748014e-05,
|
| 148 |
+
"loss": 1.2422699737548828,
|
| 149 |
+
"mean_token_accuracy": 0.6885153651237488,
|
| 150 |
+
"num_tokens": 18519114.0,
|
| 151 |
+
"step": 650
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"entropy": 1.54343731880188,
|
| 155 |
+
"epoch": 0.23809523809523808,
|
| 156 |
+
"grad_norm": 1.1484375,
|
| 157 |
+
"learning_rate": 1.58683314415437e-05,
|
| 158 |
+
"loss": 1.2279574584960937,
|
| 159 |
+
"mean_token_accuracy": 0.6913824212551117,
|
| 160 |
+
"num_tokens": 19930551.0,
|
| 161 |
+
"step": 700
|
| 162 |
+
},
|
| 163 |
+
{
|
| 164 |
+
"entropy": 1.524279227256775,
|
| 165 |
+
"epoch": 0.25510204081632654,
|
| 166 |
+
"grad_norm": 1.140625,
|
| 167 |
+
"learning_rate": 1.700340522133939e-05,
|
| 168 |
+
"loss": 1.2193751525878906,
|
| 169 |
+
"mean_token_accuracy": 0.6942444491386414,
|
| 170 |
+
"num_tokens": 21345785.0,
|
| 171 |
+
"step": 750
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"entropy": 1.527976930141449,
|
| 175 |
+
"epoch": 0.272108843537415,
|
| 176 |
+
"grad_norm": 1.1171875,
|
| 177 |
+
"learning_rate": 1.8138479001135073e-05,
|
| 178 |
+
"loss": 1.212171859741211,
|
| 179 |
+
"mean_token_accuracy": 0.6960112357139587,
|
| 180 |
+
"num_tokens": 22754238.0,
|
| 181 |
+
"step": 800
|
| 182 |
+
},
|
| 183 |
+
{
|
| 184 |
+
"entropy": 1.5164598178863526,
|
| 185 |
+
"epoch": 0.2891156462585034,
|
| 186 |
+
"grad_norm": 1.109375,
|
| 187 |
+
"learning_rate": 1.9273552780930764e-05,
|
| 188 |
+
"loss": 1.2044546508789062,
|
| 189 |
+
"mean_token_accuracy": 0.6967348265647888,
|
| 190 |
+
"num_tokens": 24167833.0,
|
| 191 |
+
"step": 850
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"entropy": 1.4968056178092957,
|
| 195 |
+
"epoch": 0.30612244897959184,
|
| 196 |
+
"grad_norm": 1.0234375,
|
| 197 |
+
"learning_rate": 1.9999746322867523e-05,
|
| 198 |
+
"loss": 1.2030063629150392,
|
| 199 |
+
"mean_token_accuracy": 0.6967223298549652,
|
| 200 |
+
"num_tokens": 25597839.0,
|
| 201 |
+
"step": 900
|
| 202 |
+
},
|
| 203 |
+
{
|
| 204 |
+
"entropy": 1.5187969017028808,
|
| 205 |
+
"epoch": 0.3231292517006803,
|
| 206 |
+
"grad_norm": 1.1796875,
|
| 207 |
+
"learning_rate": 1.9996379823325586e-05,
|
| 208 |
+
"loss": 1.2039017486572265,
|
| 209 |
+
"mean_token_accuracy": 0.69672048330307,
|
| 210 |
+
"num_tokens": 27039491.0,
|
| 211 |
+
"step": 950
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"entropy": 1.5000163841247558,
|
| 215 |
+
"epoch": 0.3401360544217687,
|
| 216 |
+
"grad_norm": 1.0546875,
|
| 217 |
+
"learning_rate": 1.9989100081556328e-05,
|
| 218 |
+
"loss": 1.2096540832519531,
|
| 219 |
+
"mean_token_accuracy": 0.696036319732666,
|
| 220 |
+
"num_tokens": 28462525.0,
|
| 221 |
+
"step": 1000
|
| 222 |
+
},
|
| 223 |
+
{
|
| 224 |
+
"epoch": 0.3401360544217687,
|
| 225 |
+
"eval_entropy": 1.4825329389892707,
|
| 226 |
+
"eval_loss": 1.1529995203018188,
|
| 227 |
+
"eval_mean_token_accuracy": 0.7072621358542883,
|
| 228 |
+
"eval_num_tokens": 28462525.0,
|
| 229 |
+
"eval_runtime": 9.42,
|
| 230 |
+
"eval_samples_per_second": 100.637,
|
| 231 |
+
"eval_steps_per_second": 12.633,
|
| 232 |
+
"step": 1000
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"entropy": 1.4857492208480836,
|
| 236 |
+
"epoch": 0.35714285714285715,
|
| 237 |
+
"grad_norm": 1.0625,
|
| 238 |
+
"learning_rate": 1.997790994733071e-05,
|
| 239 |
+
"loss": 1.1972512054443358,
|
| 240 |
+
"mean_token_accuracy": 0.6976107740402222,
|
| 241 |
+
"num_tokens": 29880725.0,
|
| 242 |
+
"step": 1050
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"entropy": 1.4679273128509522,
|
| 246 |
+
"epoch": 0.3741496598639456,
|
| 247 |
+
"grad_norm": 1.125,
|
| 248 |
+
"learning_rate": 1.9962813801205148e-05,
|
| 249 |
+
"loss": 1.1646170806884766,
|
| 250 |
+
"mean_token_accuracy": 0.705207599401474,
|
| 251 |
+
"num_tokens": 31288250.0,
|
| 252 |
+
"step": 1100
|
| 253 |
+
},
|
| 254 |
+
{
|
| 255 |
+
"entropy": 1.482924530506134,
|
| 256 |
+
"epoch": 0.391156462585034,
|
| 257 |
+
"grad_norm": 1.25,
|
| 258 |
+
"learning_rate": 1.9943817552806678e-05,
|
| 259 |
+
"loss": 1.1940898132324218,
|
| 260 |
+
"mean_token_accuracy": 0.6988638079166413,
|
| 261 |
+
"num_tokens": 32696424.0,
|
| 262 |
+
"step": 1150
|
| 263 |
+
},
|
| 264 |
+
{
|
| 265 |
+
"entropy": 1.4768511247634888,
|
| 266 |
+
"epoch": 0.40816326530612246,
|
| 267 |
+
"grad_norm": 1.1015625,
|
| 268 |
+
"learning_rate": 1.9920928638519545e-05,
|
| 269 |
+
"loss": 1.1782675933837892,
|
| 270 |
+
"mean_token_accuracy": 0.7024707806110382,
|
| 271 |
+
"num_tokens": 34115752.0,
|
| 272 |
+
"step": 1200
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"entropy": 1.4849171042442322,
|
| 276 |
+
"epoch": 0.42517006802721086,
|
| 277 |
+
"grad_norm": 1.1015625,
|
| 278 |
+
"learning_rate": 1.9894156018574104e-05,
|
| 279 |
+
"loss": 1.1730673217773437,
|
| 280 |
+
"mean_token_accuracy": 0.7031274008750915,
|
| 281 |
+
"num_tokens": 35496076.0,
|
| 282 |
+
"step": 1250
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"entropy": 1.477535102367401,
|
| 286 |
+
"epoch": 0.4421768707482993,
|
| 287 |
+
"grad_norm": 1.140625,
|
| 288 |
+
"learning_rate": 1.9863510173539193e-05,
|
| 289 |
+
"loss": 1.1849075317382813,
|
| 290 |
+
"mean_token_accuracy": 0.6999362814426422,
|
| 291 |
+
"num_tokens": 36924227.0,
|
| 292 |
+
"step": 1300
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"entropy": 1.4699360299110413,
|
| 296 |
+
"epoch": 0.45918367346938777,
|
| 297 |
+
"grad_norm": 0.9921875,
|
| 298 |
+
"learning_rate": 1.982900310021936e-05,
|
| 299 |
+
"loss": 1.1758278656005858,
|
| 300 |
+
"mean_token_accuracy": 0.7016613483428955,
|
| 301 |
+
"num_tokens": 38298695.0,
|
| 302 |
+
"step": 1350
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"entropy": 1.4796976780891418,
|
| 306 |
+
"epoch": 0.47619047619047616,
|
| 307 |
+
"grad_norm": 1.2421875,
|
| 308 |
+
"learning_rate": 1.9790648306958504e-05,
|
| 309 |
+
"loss": 1.1764393615722657,
|
| 310 |
+
"mean_token_accuracy": 0.7006231868267059,
|
| 311 |
+
"num_tokens": 39730904.0,
|
| 312 |
+
"step": 1400
|
| 313 |
+
},
|
| 314 |
+
{
|
| 315 |
+
"entropy": 1.4659403300285339,
|
| 316 |
+
"epoch": 0.4931972789115646,
|
| 317 |
+
"grad_norm": 1.109375,
|
| 318 |
+
"learning_rate": 1.974846080835184e-05,
|
| 319 |
+
"loss": 1.1739279174804687,
|
| 320 |
+
"mean_token_accuracy": 0.7042246878147125,
|
| 321 |
+
"num_tokens": 41159605.0,
|
| 322 |
+
"step": 1450
|
| 323 |
+
},
|
| 324 |
+
{
|
| 325 |
+
"entropy": 1.4660640573501587,
|
| 326 |
+
"epoch": 0.5102040816326531,
|
| 327 |
+
"grad_norm": 1.15625,
|
| 328 |
+
"learning_rate": 1.9702457119368175e-05,
|
| 329 |
+
"loss": 1.1742501831054688,
|
| 330 |
+
"mean_token_accuracy": 0.7028728783130646,
|
| 331 |
+
"num_tokens": 42591337.0,
|
| 332 |
+
"step": 1500
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 0.5102040816326531,
|
| 336 |
+
"eval_entropy": 1.4500642824573677,
|
| 337 |
+
"eval_loss": 1.1341153383255005,
|
| 338 |
+
"eval_mean_token_accuracy": 0.7110618072397569,
|
| 339 |
+
"eval_num_tokens": 42591337.0,
|
| 340 |
+
"eval_runtime": 9.205,
|
| 341 |
+
"eval_samples_per_second": 102.988,
|
| 342 |
+
"eval_steps_per_second": 12.928,
|
| 343 |
+
"step": 1500
|
| 344 |
+
},
|
| 345 |
+
{
|
| 346 |
+
"entropy": 1.4817289400100708,
|
| 347 |
+
"epoch": 0.5272108843537415,
|
| 348 |
+
"grad_norm": 1.15625,
|
| 349 |
+
"learning_rate": 1.965265524888488e-05,
|
| 350 |
+
"loss": 1.1780020904541015,
|
| 351 |
+
"mean_token_accuracy": 0.7016154682636261,
|
| 352 |
+
"num_tokens": 44005399.0,
|
| 353 |
+
"step": 1550
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"entropy": 1.4686012029647828,
|
| 357 |
+
"epoch": 0.54421768707483,
|
| 358 |
+
"grad_norm": 1.0859375,
|
| 359 |
+
"learning_rate": 1.9599074692638007e-05,
|
| 360 |
+
"loss": 1.1757928466796874,
|
| 361 |
+
"mean_token_accuracy": 0.7019401288032532,
|
| 362 |
+
"num_tokens": 45411127.0,
|
| 363 |
+
"step": 1600
|
| 364 |
+
},
|
| 365 |
+
{
|
| 366 |
+
"entropy": 1.4470409440994263,
|
| 367 |
+
"epoch": 0.5612244897959183,
|
| 368 |
+
"grad_norm": 1.234375,
|
| 369 |
+
"learning_rate": 1.9541736425590398e-05,
|
| 370 |
+
"loss": 1.1487213897705078,
|
| 371 |
+
"mean_token_accuracy": 0.708456654548645,
|
| 372 |
+
"num_tokens": 46828771.0,
|
| 373 |
+
"step": 1650
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"entropy": 1.4495110416412353,
|
| 377 |
+
"epoch": 0.5782312925170068,
|
| 378 |
+
"grad_norm": 1.1484375,
|
| 379 |
+
"learning_rate": 1.9480662893720672e-05,
|
| 380 |
+
"loss": 1.1448258972167968,
|
| 381 |
+
"mean_token_accuracy": 0.7081741154193878,
|
| 382 |
+
"num_tokens": 48222827.0,
|
| 383 |
+
"step": 1700
|
| 384 |
+
},
|
| 385 |
+
{
|
| 386 |
+
"entropy": 1.4605646800994874,
|
| 387 |
+
"epoch": 0.5952380952380952,
|
| 388 |
+
"grad_norm": 1.125,
|
| 389 |
+
"learning_rate": 1.9415878005236413e-05,
|
| 390 |
+
"loss": 1.1873452758789063,
|
| 391 |
+
"mean_token_accuracy": 0.6996394228935242,
|
| 392 |
+
"num_tokens": 49670122.0,
|
| 393 |
+
"step": 1750
|
| 394 |
+
},
|
| 395 |
+
{
|
| 396 |
+
"entropy": 1.4478923654556275,
|
| 397 |
+
"epoch": 0.6122448979591837,
|
| 398 |
+
"grad_norm": 1.1484375,
|
| 399 |
+
"learning_rate": 1.9347407121214917e-05,
|
| 400 |
+
"loss": 1.1648704528808593,
|
| 401 |
+
"mean_token_accuracy": 0.7041294479370117,
|
| 402 |
+
"num_tokens": 51095588.0,
|
| 403 |
+
"step": 1800
|
| 404 |
+
},
|
| 405 |
+
{
|
| 406 |
+
"entropy": 1.4468252277374267,
|
| 407 |
+
"epoch": 0.6292517006802721,
|
| 408 |
+
"grad_norm": 1.0703125,
|
| 409 |
+
"learning_rate": 1.927527704567517e-05,
|
| 410 |
+
"loss": 1.1468747711181642,
|
| 411 |
+
"mean_token_accuracy": 0.7075002765655518,
|
| 412 |
+
"num_tokens": 52520201.0,
|
| 413 |
+
"step": 1850
|
| 414 |
+
},
|
| 415 |
+
{
|
| 416 |
+
"entropy": 1.4675183987617493,
|
| 417 |
+
"epoch": 0.6462585034013606,
|
| 418 |
+
"grad_norm": 1.078125,
|
| 419 |
+
"learning_rate": 1.9199516015085025e-05,
|
| 420 |
+
"loss": 1.162491912841797,
|
| 421 |
+
"mean_token_accuracy": 0.7048231554031372,
|
| 422 |
+
"num_tokens": 53944548.0,
|
| 423 |
+
"step": 1900
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"entropy": 1.4320320653915406,
|
| 427 |
+
"epoch": 0.6632653061224489,
|
| 428 |
+
"grad_norm": 1.1640625,
|
| 429 |
+
"learning_rate": 1.912015368730756e-05,
|
| 430 |
+
"loss": 1.1547863006591796,
|
| 431 |
+
"mean_token_accuracy": 0.7064202272891998,
|
| 432 |
+
"num_tokens": 55352104.0,
|
| 433 |
+
"step": 1950
|
| 434 |
+
},
|
| 435 |
+
{
|
| 436 |
+
"entropy": 1.4646633243560792,
|
| 437 |
+
"epoch": 0.6802721088435374,
|
| 438 |
+
"grad_norm": 1.0625,
|
| 439 |
+
"learning_rate": 1.9037221129991045e-05,
|
| 440 |
+
"loss": 1.1570500183105468,
|
| 441 |
+
"mean_token_accuracy": 0.7064848113059997,
|
| 442 |
+
"num_tokens": 56766925.0,
|
| 443 |
+
"step": 2000
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.6802721088435374,
|
| 447 |
+
"eval_entropy": 1.4422022985810994,
|
| 448 |
+
"eval_loss": 1.1238017082214355,
|
| 449 |
+
"eval_mean_token_accuracy": 0.7135469041952566,
|
| 450 |
+
"eval_num_tokens": 56766925.0,
|
| 451 |
+
"eval_runtime": 8.3815,
|
| 452 |
+
"eval_samples_per_second": 113.106,
|
| 453 |
+
"eval_steps_per_second": 14.198,
|
| 454 |
+
"step": 2000
|
| 455 |
+
},
|
| 456 |
+
{
|
| 457 |
+
"entropy": 1.4543561935424805,
|
| 458 |
+
"epoch": 0.6972789115646258,
|
| 459 |
+
"grad_norm": 1.21875,
|
| 460 |
+
"learning_rate": 1.8950750808407016e-05,
|
| 461 |
+
"loss": 1.1518563079833983,
|
| 462 |
+
"mean_token_accuracy": 0.7068261563777923,
|
| 463 |
+
"num_tokens": 58185378.0,
|
| 464 |
+
"step": 2050
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"entropy": 1.4502357959747314,
|
| 468 |
+
"epoch": 0.7142857142857143,
|
| 469 |
+
"grad_norm": 1.0,
|
| 470 |
+
"learning_rate": 1.8860776572741214e-05,
|
| 471 |
+
"loss": 1.1643141174316407,
|
| 472 |
+
"mean_token_accuracy": 0.7048910784721375,
|
| 473 |
+
"num_tokens": 59574639.0,
|
| 474 |
+
"step": 2100
|
| 475 |
+
},
|
| 476 |
+
{
|
| 477 |
+
"entropy": 1.4497376680374146,
|
| 478 |
+
"epoch": 0.7312925170068028,
|
| 479 |
+
"grad_norm": 1.1171875,
|
| 480 |
+
"learning_rate": 1.8767333644842414e-05,
|
| 481 |
+
"loss": 1.1560958862304687,
|
| 482 |
+
"mean_token_accuracy": 0.705675152540207,
|
| 483 |
+
"num_tokens": 60979229.0,
|
| 484 |
+
"step": 2150
|
| 485 |
+
},
|
| 486 |
+
{
|
| 487 |
+
"entropy": 1.4455422043800354,
|
| 488 |
+
"epoch": 0.7482993197278912,
|
| 489 |
+
"grad_norm": 1.1015625,
|
| 490 |
+
"learning_rate": 1.8670458604434253e-05,
|
| 491 |
+
"loss": 1.1813956451416017,
|
| 492 |
+
"mean_token_accuracy": 0.6994379079341888,
|
| 493 |
+
"num_tokens": 62398557.0,
|
| 494 |
+
"step": 2200
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"entropy": 1.4504392910003663,
|
| 498 |
+
"epoch": 0.7653061224489796,
|
| 499 |
+
"grad_norm": 1.2734375,
|
| 500 |
+
"learning_rate": 1.857018937479551e-05,
|
| 501 |
+
"loss": 1.163807601928711,
|
| 502 |
+
"mean_token_accuracy": 0.7050833916664123,
|
| 503 |
+
"num_tokens": 63824693.0,
|
| 504 |
+
"step": 2250
|
| 505 |
+
},
|
| 506 |
+
{
|
| 507 |
+
"entropy": 1.4483773112297058,
|
| 508 |
+
"epoch": 0.782312925170068,
|
| 509 |
+
"grad_norm": 1.03125,
|
| 510 |
+
"learning_rate": 1.846656520791446e-05,
|
| 511 |
+
"loss": 1.1639373779296875,
|
| 512 |
+
"mean_token_accuracy": 0.7042666506767273,
|
| 513 |
+
"num_tokens": 65262483.0,
|
| 514 |
+
"step": 2300
|
| 515 |
+
},
|
| 516 |
+
{
|
| 517 |
+
"entropy": 1.4623826432228089,
|
| 518 |
+
"epoch": 0.7993197278911565,
|
| 519 |
+
"grad_norm": 1.1171875,
|
| 520 |
+
"learning_rate": 1.835962666912303e-05,
|
| 521 |
+
"loss": 1.1614888763427735,
|
| 522 |
+
"mean_token_accuracy": 0.7056748032569885,
|
| 523 |
+
"num_tokens": 66699392.0,
|
| 524 |
+
"step": 2350
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"entropy": 1.4484982657432557,
|
| 528 |
+
"epoch": 0.8163265306122449,
|
| 529 |
+
"grad_norm": 0.94921875,
|
| 530 |
+
"learning_rate": 1.8249415621216853e-05,
|
| 531 |
+
"loss": 1.164166717529297,
|
| 532 |
+
"mean_token_accuracy": 0.7041208696365356,
|
| 533 |
+
"num_tokens": 68163337.0,
|
| 534 |
+
"step": 2400
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"entropy": 1.4403035855293274,
|
| 538 |
+
"epoch": 0.8333333333333334,
|
| 539 |
+
"grad_norm": 1.171875,
|
| 540 |
+
"learning_rate": 1.8135975208067442e-05,
|
| 541 |
+
"loss": 1.1633216857910156,
|
| 542 |
+
"mean_token_accuracy": 0.7039952325820923,
|
| 543 |
+
"num_tokens": 69584260.0,
|
| 544 |
+
"step": 2450
|
| 545 |
+
},
|
| 546 |
+
{
|
| 547 |
+
"entropy": 1.4662260293960572,
|
| 548 |
+
"epoch": 0.8503401360544217,
|
| 549 |
+
"grad_norm": 1.1484375,
|
| 550 |
+
"learning_rate": 1.80193498377328e-05,
|
| 551 |
+
"loss": 1.1636495208740234,
|
| 552 |
+
"mean_token_accuracy": 0.7047837376594543,
|
| 553 |
+
"num_tokens": 70984379.0,
|
| 554 |
+
"step": 2500
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 0.8503401360544217,
|
| 558 |
+
"eval_entropy": 1.4377584517502986,
|
| 559 |
+
"eval_loss": 1.1168923377990723,
|
| 560 |
+
"eval_mean_token_accuracy": 0.7147301846191663,
|
| 561 |
+
"eval_num_tokens": 70984379.0,
|
| 562 |
+
"eval_runtime": 7.3849,
|
| 563 |
+
"eval_samples_per_second": 128.37,
|
| 564 |
+
"eval_steps_per_second": 16.114,
|
| 565 |
+
"step": 2500
|
| 566 |
+
},
|
| 567 |
+
{
|
| 568 |
+
"entropy": 1.443549563884735,
|
| 569 |
+
"epoch": 0.8673469387755102,
|
| 570 |
+
"grad_norm": 1.046875,
|
| 571 |
+
"learning_rate": 1.7899585165073182e-05,
|
| 572 |
+
"loss": 1.148828353881836,
|
| 573 |
+
"mean_token_accuracy": 0.708294941186905,
|
| 574 |
+
"num_tokens": 72388922.0,
|
| 575 |
+
"step": 2550
|
| 576 |
+
},
|
| 577 |
+
{
|
| 578 |
+
"entropy": 1.461836290359497,
|
| 579 |
+
"epoch": 0.8843537414965986,
|
| 580 |
+
"grad_norm": 1.015625,
|
| 581 |
+
"learning_rate": 1.77767280738788e-05,
|
| 582 |
+
"loss": 1.18120361328125,
|
| 583 |
+
"mean_token_accuracy": 0.7006327843666077,
|
| 584 |
+
"num_tokens": 73820763.0,
|
| 585 |
+
"step": 2600
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"entropy": 1.4670860075950622,
|
| 589 |
+
"epoch": 0.9013605442176871,
|
| 590 |
+
"grad_norm": 1.1015625,
|
| 591 |
+
"learning_rate": 1.7650826658516375e-05,
|
| 592 |
+
"loss": 1.190813751220703,
|
| 593 |
+
"mean_token_accuracy": 0.6987618601322174,
|
| 594 |
+
"num_tokens": 75254805.0,
|
| 595 |
+
"step": 2650
|
| 596 |
+
},
|
| 597 |
+
{
|
| 598 |
+
"entropy": 1.4515372467041017,
|
| 599 |
+
"epoch": 0.9183673469387755,
|
| 600 |
+
"grad_norm": 0.9140625,
|
| 601 |
+
"learning_rate": 1.7521930205101857e-05,
|
| 602 |
+
"loss": 1.170417251586914,
|
| 603 |
+
"mean_token_accuracy": 0.7032169091701508,
|
| 604 |
+
"num_tokens": 76693263.0,
|
| 605 |
+
"step": 2700
|
| 606 |
+
},
|
| 607 |
+
{
|
| 608 |
+
"entropy": 1.4385710954666138,
|
| 609 |
+
"epoch": 0.935374149659864,
|
| 610 |
+
"grad_norm": 1.0390625,
|
| 611 |
+
"learning_rate": 1.7390089172206594e-05,
|
| 612 |
+
"loss": 1.1422464752197266,
|
| 613 |
+
"mean_token_accuracy": 0.7097269940376282,
|
| 614 |
+
"num_tokens": 78106463.0,
|
| 615 |
+
"step": 2750
|
| 616 |
+
},
|
| 617 |
+
{
|
| 618 |
+
"entropy": 1.4470832967758178,
|
| 619 |
+
"epoch": 0.9523809523809523,
|
| 620 |
+
"grad_norm": 1.09375,
|
| 621 |
+
"learning_rate": 1.7255355171104492e-05,
|
| 622 |
+
"loss": 1.169587173461914,
|
| 623 |
+
"mean_token_accuracy": 0.7027925884723664,
|
| 624 |
+
"num_tokens": 79532081.0,
|
| 625 |
+
"step": 2800
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"entropy": 1.462675015926361,
|
| 629 |
+
"epoch": 0.9693877551020408,
|
| 630 |
+
"grad_norm": 1.03125,
|
| 631 |
+
"learning_rate": 1.711778094556799e-05,
|
| 632 |
+
"loss": 1.1886711883544923,
|
| 633 |
+
"mean_token_accuracy": 0.6994748151302338,
|
| 634 |
+
"num_tokens": 80999149.0,
|
| 635 |
+
"step": 2850
|
| 636 |
+
},
|
| 637 |
+
{
|
| 638 |
+
"entropy": 1.4251791214942933,
|
| 639 |
+
"epoch": 0.9863945578231292,
|
| 640 |
+
"grad_norm": 0.953125,
|
| 641 |
+
"learning_rate": 1.6977420351220656e-05,
|
| 642 |
+
"loss": 1.138572998046875,
|
| 643 |
+
"mean_token_accuracy": 0.7102291858196259,
|
| 644 |
+
"num_tokens": 82436061.0,
|
| 645 |
+
"step": 2900
|
| 646 |
+
},
|
| 647 |
+
{
|
| 648 |
+
"entropy": 1.4440323853492736,
|
| 649 |
+
"epoch": 1.0034013605442176,
|
| 650 |
+
"grad_norm": 1.171875,
|
| 651 |
+
"learning_rate": 1.683432833445453e-05,
|
| 652 |
+
"loss": 1.1582012176513672,
|
| 653 |
+
"mean_token_accuracy": 0.7026605701446533,
|
| 654 |
+
"num_tokens": 83831706.0,
|
| 655 |
+
"step": 2950
|
| 656 |
+
},
|
| 657 |
+
{
|
| 658 |
+
"entropy": 1.4328801035881042,
|
| 659 |
+
"epoch": 1.0204081632653061,
|
| 660 |
+
"grad_norm": 1.046875,
|
| 661 |
+
"learning_rate": 1.668856091092051e-05,
|
| 662 |
+
"loss": 1.1449398040771483,
|
| 663 |
+
"mean_token_accuracy": 0.7084557473659515,
|
| 664 |
+
"num_tokens": 85243788.0,
|
| 665 |
+
"step": 3000
|
| 666 |
+
},
|
| 667 |
+
{
|
| 668 |
+
"epoch": 1.0204081632653061,
|
| 669 |
+
"eval_entropy": 1.425075779442026,
|
| 670 |
+
"eval_loss": 1.112339973449707,
|
| 671 |
+
"eval_mean_token_accuracy": 0.7156877617876068,
|
| 672 |
+
"eval_num_tokens": 85243788.0,
|
| 673 |
+
"eval_runtime": 9.148,
|
| 674 |
+
"eval_samples_per_second": 103.629,
|
| 675 |
+
"eval_steps_per_second": 13.008,
|
| 676 |
+
"step": 3000
|
| 677 |
+
},
|
| 678 |
+
{
|
| 679 |
+
"entropy": 1.457193055152893,
|
| 680 |
+
"epoch": 1.0374149659863945,
|
| 681 |
+
"grad_norm": 1.0078125,
|
| 682 |
+
"learning_rate": 1.6540175143600124e-05,
|
| 683 |
+
"loss": 1.1573681640625,
|
| 684 |
+
"mean_token_accuracy": 0.706193675994873,
|
| 685 |
+
"num_tokens": 86631509.0,
|
| 686 |
+
"step": 3050
|
| 687 |
+
},
|
| 688 |
+
{
|
| 689 |
+
"entropy": 1.4502310729026795,
|
| 690 |
+
"epoch": 1.054421768707483,
|
| 691 |
+
"grad_norm": 1.125,
|
| 692 |
+
"learning_rate": 1.6389229120467304e-05,
|
| 693 |
+
"loss": 1.1431642913818358,
|
| 694 |
+
"mean_token_accuracy": 0.7098660218715668,
|
| 695 |
+
"num_tokens": 88045606.0,
|
| 696 |
+
"step": 3100
|
| 697 |
+
},
|
| 698 |
+
{
|
| 699 |
+
"entropy": 1.4366277551651,
|
| 700 |
+
"epoch": 1.0714285714285714,
|
| 701 |
+
"grad_norm": 0.9453125,
|
| 702 |
+
"learning_rate": 1.6235781931748923e-05,
|
| 703 |
+
"loss": 1.1489845275878907,
|
| 704 |
+
"mean_token_accuracy": 0.707208307981491,
|
| 705 |
+
"num_tokens": 89467532.0,
|
| 706 |
+
"step": 3150
|
| 707 |
+
},
|
| 708 |
+
{
|
| 709 |
+
"entropy": 1.4366546082496643,
|
| 710 |
+
"epoch": 1.08843537414966,
|
| 711 |
+
"grad_norm": 1.1875,
|
| 712 |
+
"learning_rate": 1.6079893646793016e-05,
|
| 713 |
+
"loss": 1.1411451721191406,
|
| 714 |
+
"mean_token_accuracy": 0.7094867730140686,
|
| 715 |
+
"num_tokens": 90884635.0,
|
| 716 |
+
"step": 3200
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"entropy": 1.451173803806305,
|
| 720 |
+
"epoch": 1.1054421768707483,
|
| 721 |
+
"grad_norm": 0.98828125,
|
| 722 |
+
"learning_rate": 1.5921625290553628e-05,
|
| 723 |
+
"loss": 1.168065185546875,
|
| 724 |
+
"mean_token_accuracy": 0.7023441231250763,
|
| 725 |
+
"num_tokens": 92294540.0,
|
| 726 |
+
"step": 3250
|
| 727 |
+
},
|
| 728 |
+
{
|
| 729 |
+
"entropy": 1.4395431089401245,
|
| 730 |
+
"epoch": 1.1224489795918366,
|
| 731 |
+
"grad_norm": 1.0703125,
|
| 732 |
+
"learning_rate": 1.5761038819701658e-05,
|
| 733 |
+
"loss": 1.1530582427978515,
|
| 734 |
+
"mean_token_accuracy": 0.7062457311153412,
|
| 735 |
+
"num_tokens": 93722473.0,
|
| 736 |
+
"step": 3300
|
| 737 |
+
},
|
| 738 |
+
{
|
| 739 |
+
"entropy": 1.4319015264511108,
|
| 740 |
+
"epoch": 1.1394557823129252,
|
| 741 |
+
"grad_norm": 1.078125,
|
| 742 |
+
"learning_rate": 1.5598197098370928e-05,
|
| 743 |
+
"loss": 1.135885009765625,
|
| 744 |
+
"mean_token_accuracy": 0.7098275601863862,
|
| 745 |
+
"num_tokens": 95130619.0,
|
| 746 |
+
"step": 3350
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"entropy": 1.4478510427474975,
|
| 750 |
+
"epoch": 1.1564625850340136,
|
| 751 |
+
"grad_norm": 1.125,
|
| 752 |
+
"learning_rate": 1.5433163873548994e-05,
|
| 753 |
+
"loss": 1.1470938110351563,
|
| 754 |
+
"mean_token_accuracy": 0.7073779857158661,
|
| 755 |
+
"num_tokens": 96554760.0,
|
| 756 |
+
"step": 3400
|
| 757 |
+
},
|
| 758 |
+
{
|
| 759 |
+
"entropy": 1.4374464821815491,
|
| 760 |
+
"epoch": 1.1734693877551021,
|
| 761 |
+
"grad_norm": 1.1328125,
|
| 762 |
+
"learning_rate": 1.5266003750122394e-05,
|
| 763 |
+
"loss": 1.1405145263671874,
|
| 764 |
+
"mean_token_accuracy": 0.7093812191486358,
|
| 765 |
+
"num_tokens": 97955334.0,
|
| 766 |
+
"step": 3450
|
| 767 |
+
},
|
| 768 |
+
{
|
| 769 |
+
"entropy": 1.4236741828918458,
|
| 770 |
+
"epoch": 1.1904761904761905,
|
| 771 |
+
"grad_norm": 1.046875,
|
| 772 |
+
"learning_rate": 1.5096782165586037e-05,
|
| 773 |
+
"loss": 1.1330892181396484,
|
| 774 |
+
"mean_token_accuracy": 0.7110612964630127,
|
| 775 |
+
"num_tokens": 99395276.0,
|
| 776 |
+
"step": 3500
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 1.1904761904761905,
|
| 780 |
+
"eval_entropy": 1.4220578530255485,
|
| 781 |
+
"eval_loss": 1.1096009016036987,
|
| 782 |
+
"eval_mean_token_accuracy": 0.7162127835409982,
|
| 783 |
+
"eval_num_tokens": 99395276.0,
|
| 784 |
+
"eval_runtime": 7.5607,
|
| 785 |
+
"eval_samples_per_second": 125.385,
|
| 786 |
+
"eval_steps_per_second": 15.739,
|
| 787 |
+
"step": 3500
|
| 788 |
+
},
|
| 789 |
+
{
|
| 790 |
+
"entropy": 1.4502542996406556,
|
| 791 |
+
"epoch": 1.2074829931972788,
|
| 792 |
+
"grad_norm": 1.109375,
|
| 793 |
+
"learning_rate": 1.4925565364426626e-05,
|
| 794 |
+
"loss": 1.1476077270507812,
|
| 795 |
+
"mean_token_accuracy": 0.7080324268341065,
|
| 796 |
+
"num_tokens": 100822779.0,
|
| 797 |
+
"step": 3550
|
| 798 |
+
},
|
| 799 |
+
{
|
| 800 |
+
"entropy": 1.4335950636863708,
|
| 801 |
+
"epoch": 1.2244897959183674,
|
| 802 |
+
"grad_norm": 1.015625,
|
| 803 |
+
"learning_rate": 1.4752420372190235e-05,
|
| 804 |
+
"loss": 1.1536802673339843,
|
| 805 |
+
"mean_token_accuracy": 0.7067959046363831,
|
| 806 |
+
"num_tokens": 102255057.0,
|
| 807 |
+
"step": 3600
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"entropy": 1.431325912475586,
|
| 811 |
+
"epoch": 1.2414965986394557,
|
| 812 |
+
"grad_norm": 1.1171875,
|
| 813 |
+
"learning_rate": 1.4577414969244052e-05,
|
| 814 |
+
"loss": 1.1357704162597657,
|
| 815 |
+
"mean_token_accuracy": 0.7090205514431,
|
| 816 |
+
"num_tokens": 103682059.0,
|
| 817 |
+
"step": 3650
|
| 818 |
+
},
|
| 819 |
+
{
|
| 820 |
+
"entropy": 1.430570056438446,
|
| 821 |
+
"epoch": 1.2585034013605443,
|
| 822 |
+
"grad_norm": 1.0625,
|
| 823 |
+
"learning_rate": 1.4400617664242706e-05,
|
| 824 |
+
"loss": 1.1330258178710937,
|
| 825 |
+
"mean_token_accuracy": 0.7101845383644104,
|
| 826 |
+
"num_tokens": 105094949.0,
|
| 827 |
+
"step": 3700
|
| 828 |
+
},
|
| 829 |
+
{
|
| 830 |
+
"entropy": 1.4420754599571228,
|
| 831 |
+
"epoch": 1.2755102040816326,
|
| 832 |
+
"grad_norm": 0.96875,
|
| 833 |
+
"learning_rate": 1.422209766730945e-05,
|
| 834 |
+
"loss": 1.1578455352783203,
|
| 835 |
+
"mean_token_accuracy": 0.7062867367267609,
|
| 836 |
+
"num_tokens": 106525361.0,
|
| 837 |
+
"step": 3750
|
| 838 |
+
},
|
| 839 |
+
{
|
| 840 |
+
"entropy": 1.4218295288085938,
|
| 841 |
+
"epoch": 1.2925170068027212,
|
| 842 |
+
"grad_norm": 0.8984375,
|
| 843 |
+
"learning_rate": 1.4041924862942744e-05,
|
| 844 |
+
"loss": 1.1394439697265626,
|
| 845 |
+
"mean_token_accuracy": 0.7095408570766449,
|
| 846 |
+
"num_tokens": 107938288.0,
|
| 847 |
+
"step": 3800
|
| 848 |
+
},
|
| 849 |
+
{
|
| 850 |
+
"entropy": 1.425144808292389,
|
| 851 |
+
"epoch": 1.3095238095238095,
|
| 852 |
+
"grad_norm": 1.1875,
|
| 853 |
+
"learning_rate": 1.3860169782658863e-05,
|
| 854 |
+
"loss": 1.1289046478271485,
|
| 855 |
+
"mean_token_accuracy": 0.7117006468772888,
|
| 856 |
+
"num_tokens": 109343324.0,
|
| 857 |
+
"step": 3850
|
| 858 |
+
},
|
| 859 |
+
{
|
| 860 |
+
"entropy": 1.4213409280776979,
|
| 861 |
+
"epoch": 1.3265306122448979,
|
| 862 |
+
"grad_norm": 1.0625,
|
| 863 |
+
"learning_rate": 1.3676903577381231e-05,
|
| 864 |
+
"loss": 1.124376220703125,
|
| 865 |
+
"mean_token_accuracy": 0.7127799916267396,
|
| 866 |
+
"num_tokens": 110773342.0,
|
| 867 |
+
"step": 3900
|
| 868 |
+
},
|
| 869 |
+
{
|
| 870 |
+
"entropy": 1.430138614177704,
|
| 871 |
+
"epoch": 1.3435374149659864,
|
| 872 |
+
"grad_norm": 1.03125,
|
| 873 |
+
"learning_rate": 1.3492197989587221e-05,
|
| 874 |
+
"loss": 1.1303221893310547,
|
| 875 |
+
"mean_token_accuracy": 0.7113751316070557,
|
| 876 |
+
"num_tokens": 112223937.0,
|
| 877 |
+
"step": 3950
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"entropy": 1.421439278125763,
|
| 881 |
+
"epoch": 1.3605442176870748,
|
| 882 |
+
"grad_norm": 1.0390625,
|
| 883 |
+
"learning_rate": 1.3306125325223433e-05,
|
| 884 |
+
"loss": 1.128666305541992,
|
| 885 |
+
"mean_token_accuracy": 0.711964373588562,
|
| 886 |
+
"num_tokens": 113638897.0,
|
| 887 |
+
"step": 4000
|
| 888 |
+
},
|
| 889 |
+
{
|
| 890 |
+
"epoch": 1.3605442176870748,
|
| 891 |
+
"eval_entropy": 1.4151852511558212,
|
| 892 |
+
"eval_loss": 1.107796311378479,
|
| 893 |
+
"eval_mean_token_accuracy": 0.7166504459220822,
|
| 894 |
+
"eval_num_tokens": 113638897.0,
|
| 895 |
+
"eval_runtime": 8.7246,
|
| 896 |
+
"eval_samples_per_second": 108.658,
|
| 897 |
+
"eval_steps_per_second": 13.64,
|
| 898 |
+
"step": 4000
|
| 899 |
+
},
|
| 900 |
+
{
|
| 901 |
+
"entropy": 1.422073357105255,
|
| 902 |
+
"epoch": 1.3775510204081631,
|
| 903 |
+
"grad_norm": 1.1015625,
|
| 904 |
+
"learning_rate": 1.31187584254004e-05,
|
| 905 |
+
"loss": 1.137821807861328,
|
| 906 |
+
"mean_token_accuracy": 0.7102219009399414,
|
| 907 |
+
"num_tokens": 115030025.0,
|
| 908 |
+
"step": 4050
|
| 909 |
+
},
|
| 910 |
+
{
|
| 911 |
+
"entropy": 1.4335832858085633,
|
| 912 |
+
"epoch": 1.3945578231292517,
|
| 913 |
+
"grad_norm": 1.1171875,
|
| 914 |
+
"learning_rate": 1.2930170637877741e-05,
|
| 915 |
+
"loss": 1.1326773834228516,
|
| 916 |
+
"mean_token_accuracy": 0.7117250227928161,
|
| 917 |
+
"num_tokens": 116464440.0,
|
| 918 |
+
"step": 4100
|
| 919 |
+
},
|
| 920 |
+
{
|
| 921 |
+
"entropy": 1.432119870185852,
|
| 922 |
+
"epoch": 1.4115646258503403,
|
| 923 |
+
"grad_norm": 1.0546875,
|
| 924 |
+
"learning_rate": 1.2740435788351031e-05,
|
| 925 |
+
"loss": 1.144314727783203,
|
| 926 |
+
"mean_token_accuracy": 0.7074739003181457,
|
| 927 |
+
"num_tokens": 117895570.0,
|
| 928 |
+
"step": 4150
|
| 929 |
+
},
|
| 930 |
+
{
|
| 931 |
+
"entropy": 1.4382991909980773,
|
| 932 |
+
"epoch": 1.4285714285714286,
|
| 933 |
+
"grad_norm": 1.0546875,
|
| 934 |
+
"learning_rate": 1.254962815155155e-05,
|
| 935 |
+
"loss": 1.1475949859619141,
|
| 936 |
+
"mean_token_accuracy": 0.7079149842262268,
|
| 937 |
+
"num_tokens": 119315574.0,
|
| 938 |
+
"step": 4200
|
| 939 |
+
},
|
| 940 |
+
{
|
| 941 |
+
"entropy": 1.3948691391944885,
|
| 942 |
+
"epoch": 1.445578231292517,
|
| 943 |
+
"grad_norm": 0.99609375,
|
| 944 |
+
"learning_rate": 1.2357822422170266e-05,
|
| 945 |
+
"loss": 1.1130772399902344,
|
| 946 |
+
"mean_token_accuracy": 0.7150168895721436,
|
| 947 |
+
"num_tokens": 120717868.0,
|
| 948 |
+
"step": 4250
|
| 949 |
+
},
|
| 950 |
+
{
|
| 951 |
+
"entropy": 1.4280868649482727,
|
| 952 |
+
"epoch": 1.4625850340136055,
|
| 953 |
+
"grad_norm": 1.015625,
|
| 954 |
+
"learning_rate": 1.2165093685617377e-05,
|
| 955 |
+
"loss": 1.1200074005126952,
|
| 956 |
+
"mean_token_accuracy": 0.7135682916641235,
|
| 957 |
+
"num_tokens": 122146924.0,
|
| 958 |
+
"step": 4300
|
| 959 |
+
},
|
| 960 |
+
{
|
| 961 |
+
"entropy": 1.4316187024116516,
|
| 962 |
+
"epoch": 1.4795918367346939,
|
| 963 |
+
"grad_norm": 1.0390625,
|
| 964 |
+
"learning_rate": 1.1971517388628972e-05,
|
| 965 |
+
"loss": 1.1323360443115233,
|
| 966 |
+
"mean_token_accuracy": 0.7125037395954132,
|
| 967 |
+
"num_tokens": 123574594.0,
|
| 968 |
+
"step": 4350
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"entropy": 1.4233829760551453,
|
| 972 |
+
"epoch": 1.4965986394557822,
|
| 973 |
+
"grad_norm": 1.0625,
|
| 974 |
+
"learning_rate": 1.1777169309732178e-05,
|
| 975 |
+
"loss": 1.144235610961914,
|
| 976 |
+
"mean_token_accuracy": 0.707960147857666,
|
| 977 |
+
"num_tokens": 124985723.0,
|
| 978 |
+
"step": 4400
|
| 979 |
+
},
|
| 980 |
+
{
|
| 981 |
+
"entropy": 1.446813395023346,
|
| 982 |
+
"epoch": 1.5136054421768708,
|
| 983 |
+
"grad_norm": 0.9765625,
|
| 984 |
+
"learning_rate": 1.1582125529580437e-05,
|
| 985 |
+
"loss": 1.1620960235595703,
|
| 986 |
+
"mean_token_accuracy": 0.7044985663890838,
|
| 987 |
+
"num_tokens": 126444415.0,
|
| 988 |
+
"step": 4450
|
| 989 |
+
},
|
| 990 |
+
{
|
| 991 |
+
"entropy": 1.4292146420478822,
|
| 992 |
+
"epoch": 1.5306122448979593,
|
| 993 |
+
"grad_norm": 1.1328125,
|
| 994 |
+
"learning_rate": 1.1386462401170562e-05,
|
| 995 |
+
"loss": 1.1338434600830078,
|
| 996 |
+
"mean_token_accuracy": 0.7091768038272858,
|
| 997 |
+
"num_tokens": 127870485.0,
|
| 998 |
+
"step": 4500
|
| 999 |
+
},
|
| 1000 |
+
{
|
| 1001 |
+
"epoch": 1.5306122448979593,
|
| 1002 |
+
"eval_entropy": 1.4135069837089347,
|
| 1003 |
+
"eval_loss": 1.106415033340454,
|
| 1004 |
+
"eval_mean_token_accuracy": 0.7169664757592338,
|
| 1005 |
+
"eval_num_tokens": 127870485.0,
|
| 1006 |
+
"eval_runtime": 8.6098,
|
| 1007 |
+
"eval_samples_per_second": 110.107,
|
| 1008 |
+
"eval_steps_per_second": 13.821,
|
| 1009 |
+
"step": 4500
|
| 1010 |
+
},
|
| 1011 |
+
{
|
| 1012 |
+
"entropy": 1.412301104068756,
|
| 1013 |
+
"epoch": 1.5476190476190477,
|
| 1014 |
+
"grad_norm": 1.1171875,
|
| 1015 |
+
"learning_rate": 1.1190256519953097e-05,
|
| 1016 |
+
"loss": 1.1280068206787108,
|
| 1017 |
+
"mean_token_accuracy": 0.7108085489273072,
|
| 1018 |
+
"num_tokens": 129318094.0,
|
| 1019 |
+
"step": 4550
|
| 1020 |
+
},
|
| 1021 |
+
{
|
| 1022 |
+
"entropy": 1.4223233222961427,
|
| 1023 |
+
"epoch": 1.564625850340136,
|
| 1024 |
+
"grad_norm": 1.1484375,
|
| 1025 |
+
"learning_rate": 1.0993584693847827e-05,
|
| 1026 |
+
"loss": 1.1158720397949218,
|
| 1027 |
+
"mean_token_accuracy": 0.7139767467975616,
|
| 1028 |
+
"num_tokens": 130726607.0,
|
| 1029 |
+
"step": 4600
|
| 1030 |
+
},
|
| 1031 |
+
{
|
| 1032 |
+
"entropy": 1.42511239528656,
|
| 1033 |
+
"epoch": 1.5816326530612246,
|
| 1034 |
+
"grad_norm": 1.1171875,
|
| 1035 |
+
"learning_rate": 1.0796523913176092e-05,
|
| 1036 |
+
"loss": 1.1147970581054687,
|
| 1037 |
+
"mean_token_accuracy": 0.7152622473239899,
|
| 1038 |
+
"num_tokens": 132145387.0,
|
| 1039 |
+
"step": 4650
|
| 1040 |
+
},
|
| 1041 |
+
{
|
| 1042 |
+
"entropy": 1.4350685811042785,
|
| 1043 |
+
"epoch": 1.598639455782313,
|
| 1044 |
+
"grad_norm": 1.0625,
|
| 1045 |
+
"learning_rate": 1.0599151320521703e-05,
|
| 1046 |
+
"loss": 1.1316419982910155,
|
| 1047 |
+
"mean_token_accuracy": 0.7108878707885742,
|
| 1048 |
+
"num_tokens": 133560830.0,
|
| 1049 |
+
"step": 4700
|
| 1050 |
+
},
|
| 1051 |
+
{
|
| 1052 |
+
"entropy": 1.4368190813064574,
|
| 1053 |
+
"epoch": 1.6156462585034013,
|
| 1054 |
+
"grad_norm": 1.0703125,
|
| 1055 |
+
"learning_rate": 1.040154418053223e-05,
|
| 1056 |
+
"loss": 1.1656169128417968,
|
| 1057 |
+
"mean_token_accuracy": 0.7034963488578796,
|
| 1058 |
+
"num_tokens": 134992182.0,
|
| 1059 |
+
"step": 4750
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"entropy": 1.4219933152198792,
|
| 1063 |
+
"epoch": 1.6326530612244898,
|
| 1064 |
+
"grad_norm": 1.0234375,
|
| 1065 |
+
"learning_rate": 1.0203779849672546e-05,
|
| 1066 |
+
"loss": 1.1332920837402343,
|
| 1067 |
+
"mean_token_accuracy": 0.7100002634525299,
|
| 1068 |
+
"num_tokens": 136426899.0,
|
| 1069 |
+
"step": 4800
|
| 1070 |
+
},
|
| 1071 |
+
{
|
| 1072 |
+
"entropy": 1.4121305012702943,
|
| 1073 |
+
"epoch": 1.6496598639455784,
|
| 1074 |
+
"grad_norm": 1.1484375,
|
| 1075 |
+
"learning_rate": 1.0005935745942395e-05,
|
| 1076 |
+
"loss": 1.120484619140625,
|
| 1077 |
+
"mean_token_accuracy": 0.7137669050693511,
|
| 1078 |
+
"num_tokens": 137821124.0,
|
| 1079 |
+
"step": 4850
|
| 1080 |
+
},
|
| 1081 |
+
{
|
| 1082 |
+
"entropy": 1.4223187255859375,
|
| 1083 |
+
"epoch": 1.6666666666666665,
|
| 1084 |
+
"grad_norm": 1.21875,
|
| 1085 |
+
"learning_rate": 9.808089318569872e-06,
|
| 1086 |
+
"loss": 1.13340087890625,
|
| 1087 |
+
"mean_token_accuracy": 0.7105760216712952,
|
| 1088 |
+
"num_tokens": 139238831.0,
|
| 1089 |
+
"step": 4900
|
| 1090 |
+
},
|
| 1091 |
+
{
|
| 1092 |
+
"entropy": 1.4339079999923705,
|
| 1093 |
+
"epoch": 1.683673469387755,
|
| 1094 |
+
"grad_norm": 1.1171875,
|
| 1095 |
+
"learning_rate": 9.610318017692717e-06,
|
| 1096 |
+
"loss": 1.131591567993164,
|
| 1097 |
+
"mean_token_accuracy": 0.7117097544670105,
|
| 1098 |
+
"num_tokens": 140670083.0,
|
| 1099 |
+
"step": 4950
|
| 1100 |
+
},
|
| 1101 |
+
{
|
| 1102 |
+
"entropy": 1.42487220287323,
|
| 1103 |
+
"epoch": 1.7006802721088436,
|
| 1104 |
+
"grad_norm": 1.0,
|
| 1105 |
+
"learning_rate": 9.412699264039187e-06,
|
| 1106 |
+
"loss": 1.1361915588378906,
|
| 1107 |
+
"mean_token_accuracy": 0.709000414609909,
|
| 1108 |
+
"num_tokens": 142109640.0,
|
| 1109 |
+
"step": 5000
|
| 1110 |
+
},
|
| 1111 |
+
{
|
| 1112 |
+
"epoch": 1.7006802721088436,
|
| 1113 |
+
"eval_entropy": 1.4130272725049187,
|
| 1114 |
+
"eval_loss": 1.105644941329956,
|
| 1115 |
+
"eval_mean_token_accuracy": 0.7170687387971317,
|
| 1116 |
+
"eval_num_tokens": 142109640.0,
|
| 1117 |
+
"eval_runtime": 9.1754,
|
| 1118 |
+
"eval_samples_per_second": 103.32,
|
| 1119 |
+
"eval_steps_per_second": 12.97,
|
| 1120 |
+
"step": 5000
|
| 1121 |
+
},
|
| 1122 |
+
{
|
| 1123 |
+
"entropy": 1.43812607049942,
|
| 1124 |
+
"epoch": 1.717687074829932,
|
| 1125 |
+
"grad_norm": 1.109375,
|
| 1126 |
+
"learning_rate": 9.215310418620513e-06,
|
| 1127 |
+
"loss": 1.1505773162841797,
|
| 1128 |
+
"mean_token_accuracy": 0.7071945500373841,
|
| 1129 |
+
"num_tokens": 143529976.0,
|
| 1130 |
+
"step": 5050
|
| 1131 |
+
},
|
| 1132 |
+
{
|
| 1133 |
+
"entropy": 1.4412685680389403,
|
| 1134 |
+
"epoch": 1.7346938775510203,
|
| 1135 |
+
"grad_norm": 1.2890625,
|
| 1136 |
+
"learning_rate": 9.018228752446671e-06,
|
| 1137 |
+
"loss": 1.1481983947753907,
|
| 1138 |
+
"mean_token_accuracy": 0.7087244749069214,
|
| 1139 |
+
"num_tokens": 144937854.0,
|
| 1140 |
+
"step": 5100
|
| 1141 |
+
},
|
| 1142 |
+
{
|
| 1143 |
+
"entropy": 1.4139077353477478,
|
| 1144 |
+
"epoch": 1.751700680272109,
|
| 1145 |
+
"grad_norm": 1.046875,
|
| 1146 |
+
"learning_rate": 8.821531416277427e-06,
|
| 1147 |
+
"loss": 1.1202762603759766,
|
| 1148 |
+
"mean_token_accuracy": 0.7141082406044006,
|
| 1149 |
+
"num_tokens": 146358885.0,
|
| 1150 |
+
"step": 5150
|
| 1151 |
+
},
|
| 1152 |
+
{
|
| 1153 |
+
"entropy": 1.4265229201316834,
|
| 1154 |
+
"epoch": 1.7687074829931972,
|
| 1155 |
+
"grad_norm": 1.0390625,
|
| 1156 |
+
"learning_rate": 8.625295410420451e-06,
|
| 1157 |
+
"loss": 1.1334566497802734,
|
| 1158 |
+
"mean_token_accuracy": 0.7119065308570862,
|
| 1159 |
+
"num_tokens": 147797164.0,
|
| 1160 |
+
"step": 5200
|
| 1161 |
+
},
|
| 1162 |
+
{
|
| 1163 |
+
"entropy": 1.4280506682395935,
|
| 1164 |
+
"epoch": 1.7857142857142856,
|
| 1165 |
+
"grad_norm": 1.125,
|
| 1166 |
+
"learning_rate": 8.429597554588291e-06,
|
| 1167 |
+
"loss": 1.1369068908691407,
|
| 1168 |
+
"mean_token_accuracy": 0.7097280561923981,
|
| 1169 |
+
"num_tokens": 149201547.0,
|
| 1170 |
+
"step": 5250
|
| 1171 |
+
},
|
| 1172 |
+
{
|
| 1173 |
+
"entropy": 1.4155354285240174,
|
| 1174 |
+
"epoch": 1.8027210884353742,
|
| 1175 |
+
"grad_norm": 1.0390625,
|
| 1176 |
+
"learning_rate": 8.234514457826101e-06,
|
| 1177 |
+
"loss": 1.1342711639404297,
|
| 1178 |
+
"mean_token_accuracy": 0.7102336215972901,
|
| 1179 |
+
"num_tokens": 150625894.0,
|
| 1180 |
+
"step": 5300
|
| 1181 |
+
},
|
| 1182 |
+
{
|
| 1183 |
+
"entropy": 1.4110008382797241,
|
| 1184 |
+
"epoch": 1.8197278911564627,
|
| 1185 |
+
"grad_norm": 1.1640625,
|
| 1186 |
+
"learning_rate": 8.040122488521802e-06,
|
| 1187 |
+
"loss": 1.1284471893310546,
|
| 1188 |
+
"mean_token_accuracy": 0.7116459333896636,
|
| 1189 |
+
"num_tokens": 152068477.0,
|
| 1190 |
+
"step": 5350
|
| 1191 |
+
},
|
| 1192 |
+
{
|
| 1193 |
+
"entropy": 1.432766444683075,
|
| 1194 |
+
"epoch": 1.836734693877551,
|
| 1195 |
+
"grad_norm": 1.109375,
|
| 1196 |
+
"learning_rate": 7.84649774451044e-06,
|
| 1197 |
+
"loss": 1.1442298126220702,
|
| 1198 |
+
"mean_token_accuracy": 0.7084469604492187,
|
| 1199 |
+
"num_tokens": 153452805.0,
|
| 1200 |
+
"step": 5400
|
| 1201 |
+
},
|
| 1202 |
+
{
|
| 1203 |
+
"entropy": 1.4228289747238159,
|
| 1204 |
+
"epoch": 1.8537414965986394,
|
| 1205 |
+
"grad_norm": 1.0,
|
| 1206 |
+
"learning_rate": 7.653716023284498e-06,
|
| 1207 |
+
"loss": 1.1224955749511718,
|
| 1208 |
+
"mean_token_accuracy": 0.7124292123317718,
|
| 1209 |
+
"num_tokens": 154873137.0,
|
| 1210 |
+
"step": 5450
|
| 1211 |
+
},
|
| 1212 |
+
{
|
| 1213 |
+
"entropy": 1.4096328234672546,
|
| 1214 |
+
"epoch": 1.870748299319728,
|
| 1215 |
+
"grad_norm": 1.109375,
|
| 1216 |
+
"learning_rate": 7.46185279232176e-06,
|
| 1217 |
+
"loss": 1.1197357940673829,
|
| 1218 |
+
"mean_token_accuracy": 0.7131345391273498,
|
| 1219 |
+
"num_tokens": 156324700.0,
|
| 1220 |
+
"step": 5500
|
| 1221 |
+
},
|
| 1222 |
+
{
|
| 1223 |
+
"epoch": 1.870748299319728,
|
| 1224 |
+
"eval_entropy": 1.4128466764418017,
|
| 1225 |
+
"eval_loss": 1.1052268743515015,
|
| 1226 |
+
"eval_mean_token_accuracy": 0.7171510763528968,
|
| 1227 |
+
"eval_num_tokens": 156324700.0,
|
| 1228 |
+
"eval_runtime": 9.0481,
|
| 1229 |
+
"eval_samples_per_second": 104.773,
|
| 1230 |
+
"eval_steps_per_second": 13.152,
|
| 1231 |
+
"step": 5500
|
| 1232 |
+
},
|
| 1233 |
+
{
|
| 1234 |
+
"entropy": 1.4428221464157105,
|
| 1235 |
+
"epoch": 1.8877551020408163,
|
| 1236 |
+
"grad_norm": 0.96484375,
|
| 1237 |
+
"learning_rate": 7.27098315954234e-06,
|
| 1238 |
+
"loss": 1.1470510101318359,
|
| 1239 |
+
"mean_token_accuracy": 0.7073501777648926,
|
| 1240 |
+
"num_tokens": 157737046.0,
|
| 1241 |
+
"step": 5550
|
| 1242 |
+
},
|
| 1243 |
+
{
|
| 1244 |
+
"entropy": 1.4279747819900512,
|
| 1245 |
+
"epoch": 1.9047619047619047,
|
| 1246 |
+
"grad_norm": 1.0390625,
|
| 1247 |
+
"learning_rate": 7.081181843906513e-06,
|
| 1248 |
+
"loss": 1.1433404541015626,
|
| 1249 |
+
"mean_token_accuracy": 0.7082498788833618,
|
| 1250 |
+
"num_tokens": 159172992.0,
|
| 1251 |
+
"step": 5600
|
| 1252 |
+
},
|
| 1253 |
+
{
|
| 1254 |
+
"entropy": 1.4385703134536743,
|
| 1255 |
+
"epoch": 1.9217687074829932,
|
| 1256 |
+
"grad_norm": 1.1328125,
|
| 1257 |
+
"learning_rate": 6.892523146164782e-06,
|
| 1258 |
+
"loss": 1.1372301483154297,
|
| 1259 |
+
"mean_token_accuracy": 0.7099990272521972,
|
| 1260 |
+
"num_tokens": 160577822.0,
|
| 1261 |
+
"step": 5650
|
| 1262 |
+
},
|
| 1263 |
+
{
|
| 1264 |
+
"entropy": 1.4102177619934082,
|
| 1265 |
+
"epoch": 1.9387755102040818,
|
| 1266 |
+
"grad_norm": 1.0625,
|
| 1267 |
+
"learning_rate": 6.705080919771655e-06,
|
| 1268 |
+
"loss": 1.1329367065429687,
|
| 1269 |
+
"mean_token_accuracy": 0.7119775557518006,
|
| 1270 |
+
"num_tokens": 161991010.0,
|
| 1271 |
+
"step": 5700
|
| 1272 |
+
},
|
| 1273 |
+
{
|
| 1274 |
+
"entropy": 1.4429755330085754,
|
| 1275 |
+
"epoch": 1.95578231292517,
|
| 1276 |
+
"grad_norm": 1.125,
|
| 1277 |
+
"learning_rate": 6.51892854197454e-06,
|
| 1278 |
+
"loss": 1.1424610900878907,
|
| 1279 |
+
"mean_token_accuracy": 0.7083716082572937,
|
| 1280 |
+
"num_tokens": 163405551.0,
|
| 1281 |
+
"step": 5750
|
| 1282 |
+
},
|
| 1283 |
+
{
|
| 1284 |
+
"entropy": 1.4363023614883423,
|
| 1285 |
+
"epoch": 1.9727891156462585,
|
| 1286 |
+
"grad_norm": 1.109375,
|
| 1287 |
+
"learning_rate": 6.334138885089061e-06,
|
| 1288 |
+
"loss": 1.142376937866211,
|
| 1289 |
+
"mean_token_accuracy": 0.709513326883316,
|
| 1290 |
+
"num_tokens": 164828847.0,
|
| 1291 |
+
"step": 5800
|
| 1292 |
+
},
|
| 1293 |
+
{
|
| 1294 |
+
"entropy": 1.435585732460022,
|
| 1295 |
+
"epoch": 1.989795918367347,
|
| 1296 |
+
"grad_norm": 1.0234375,
|
| 1297 |
+
"learning_rate": 6.150784287972005e-06,
|
| 1298 |
+
"loss": 1.157130126953125,
|
| 1299 |
+
"mean_token_accuracy": 0.7060146403312683,
|
| 1300 |
+
"num_tokens": 166264927.0,
|
| 1301 |
+
"step": 5850
|
| 1302 |
+
},
|
| 1303 |
+
{
|
| 1304 |
+
"entropy": 1.4074757862091065,
|
| 1305 |
+
"epoch": 2.006802721088435,
|
| 1306 |
+
"grad_norm": 1.0546875,
|
| 1307 |
+
"learning_rate": 5.968936527703144e-06,
|
| 1308 |
+
"loss": 1.1172750091552734,
|
| 1309 |
+
"mean_token_accuracy": 0.7146967911720276,
|
| 1310 |
+
"num_tokens": 167673391.0,
|
| 1311 |
+
"step": 5900
|
| 1312 |
+
},
|
| 1313 |
+
{
|
| 1314 |
+
"entropy": 1.4369916796684266,
|
| 1315 |
+
"epoch": 2.0238095238095237,
|
| 1316 |
+
"grad_norm": 0.98828125,
|
| 1317 |
+
"learning_rate": 5.78866679148695e-06,
|
| 1318 |
+
"loss": 1.1662197875976563,
|
| 1319 |
+
"mean_token_accuracy": 0.7031553852558136,
|
| 1320 |
+
"num_tokens": 169094709.0,
|
| 1321 |
+
"step": 5950
|
| 1322 |
+
},
|
| 1323 |
+
{
|
| 1324 |
+
"entropy": 1.4330995082855225,
|
| 1325 |
+
"epoch": 2.0408163265306123,
|
| 1326 |
+
"grad_norm": 1.03125,
|
| 1327 |
+
"learning_rate": 5.610045648785225e-06,
|
| 1328 |
+
"loss": 1.1259635162353516,
|
| 1329 |
+
"mean_token_accuracy": 0.7129771506786347,
|
| 1330 |
+
"num_tokens": 170523291.0,
|
| 1331 |
+
"step": 6000
|
| 1332 |
+
},
|
| 1333 |
+
{
|
| 1334 |
+
"epoch": 2.0408163265306123,
|
| 1335 |
+
"eval_entropy": 1.4114271442429358,
|
| 1336 |
+
"eval_loss": 1.1050275564193726,
|
| 1337 |
+
"eval_mean_token_accuracy": 0.7171849367999229,
|
| 1338 |
+
"eval_num_tokens": 170523291.0,
|
| 1339 |
+
"eval_runtime": 9.0969,
|
| 1340 |
+
"eval_samples_per_second": 104.211,
|
| 1341 |
+
"eval_steps_per_second": 13.081,
|
| 1342 |
+
"step": 6000
|
| 1343 |
+
},
|
| 1344 |
+
{
|
| 1345 |
+
"entropy": 1.4260020923614503,
|
| 1346 |
+
"epoch": 2.057823129251701,
|
| 1347 |
+
"grad_norm": 1.078125,
|
| 1348 |
+
"learning_rate": 5.433143023691547e-06,
|
| 1349 |
+
"loss": 1.1178092193603515,
|
| 1350 |
+
"mean_token_accuracy": 0.7144936168193817,
|
| 1351 |
+
"num_tokens": 171938887.0,
|
| 1352 |
+
"step": 6050
|
| 1353 |
+
},
|
| 1354 |
+
{
|
| 1355 |
+
"entropy": 1.4517771482467652,
|
| 1356 |
+
"epoch": 2.074829931972789,
|
| 1357 |
+
"grad_norm": 1.0703125,
|
| 1358 |
+
"learning_rate": 5.2580281675584045e-06,
|
| 1359 |
+
"loss": 1.1548236083984376,
|
| 1360 |
+
"mean_token_accuracy": 0.7052713608741761,
|
| 1361 |
+
"num_tokens": 173371802.0,
|
| 1362 |
+
"step": 6100
|
| 1363 |
+
},
|
| 1364 |
+
{
|
| 1365 |
+
"entropy": 1.41314546585083,
|
| 1366 |
+
"epoch": 2.0918367346938775,
|
| 1367 |
+
"grad_norm": 1.0625,
|
| 1368 |
+
"learning_rate": 5.084769631887611e-06,
|
| 1369 |
+
"loss": 1.116751708984375,
|
| 1370 |
+
"mean_token_accuracy": 0.7143128430843353,
|
| 1371 |
+
"num_tokens": 174783437.0,
|
| 1372 |
+
"step": 6150
|
| 1373 |
+
},
|
| 1374 |
+
{
|
| 1375 |
+
"entropy": 1.4380582642555237,
|
| 1376 |
+
"epoch": 2.108843537414966,
|
| 1377 |
+
"grad_norm": 1.109375,
|
| 1378 |
+
"learning_rate": 4.913435241494749e-06,
|
| 1379 |
+
"loss": 1.14319091796875,
|
| 1380 |
+
"mean_token_accuracy": 0.7082942843437194,
|
| 1381 |
+
"num_tokens": 176235112.0,
|
| 1382 |
+
"step": 6200
|
| 1383 |
+
},
|
| 1384 |
+
{
|
| 1385 |
+
"entropy": 1.432828288078308,
|
| 1386 |
+
"epoch": 2.1258503401360542,
|
| 1387 |
+
"grad_norm": 1.046875,
|
| 1388 |
+
"learning_rate": 4.744092067958067e-06,
|
| 1389 |
+
"loss": 1.1341230010986327,
|
| 1390 |
+
"mean_token_accuracy": 0.7104381382465362,
|
| 1391 |
+
"num_tokens": 177657881.0,
|
| 1392 |
+
"step": 6250
|
| 1393 |
+
},
|
| 1394 |
+
{
|
| 1395 |
+
"entropy": 1.4256313157081604,
|
| 1396 |
+
"epoch": 2.142857142857143,
|
| 1397 |
+
"grad_norm": 1.0078125,
|
| 1398 |
+
"learning_rate": 4.5768064033622315e-06,
|
| 1399 |
+
"loss": 1.13789306640625,
|
| 1400 |
+
"mean_token_accuracy": 0.7089906525611878,
|
| 1401 |
+
"num_tokens": 179083511.0,
|
| 1402 |
+
"step": 6300
|
| 1403 |
+
},
|
| 1404 |
+
{
|
| 1405 |
+
"entropy": 1.4250177597999574,
|
| 1406 |
+
"epoch": 2.1598639455782314,
|
| 1407 |
+
"grad_norm": 1.2421875,
|
| 1408 |
+
"learning_rate": 4.411643734347228e-06,
|
| 1409 |
+
"loss": 1.142816925048828,
|
| 1410 |
+
"mean_token_accuracy": 0.7079160177707672,
|
| 1411 |
+
"num_tokens": 180486838.0,
|
| 1412 |
+
"step": 6350
|
| 1413 |
+
},
|
| 1414 |
+
{
|
| 1415 |
+
"entropy": 1.397556071281433,
|
| 1416 |
+
"epoch": 2.17687074829932,
|
| 1417 |
+
"grad_norm": 0.9921875,
|
| 1418 |
+
"learning_rate": 4.248668716472597e-06,
|
| 1419 |
+
"loss": 1.1062926483154296,
|
| 1420 |
+
"mean_token_accuracy": 0.716565626859665,
|
| 1421 |
+
"num_tokens": 181896305.0,
|
| 1422 |
+
"step": 6400
|
| 1423 |
+
},
|
| 1424 |
+
{
|
| 1425 |
+
"entropy": 1.4365149688720704,
|
| 1426 |
+
"epoch": 2.193877551020408,
|
| 1427 |
+
"grad_norm": 0.9609375,
|
| 1428 |
+
"learning_rate": 4.087945148906956e-06,
|
| 1429 |
+
"loss": 1.1448672485351563,
|
| 1430 |
+
"mean_token_accuracy": 0.7083494472503662,
|
| 1431 |
+
"num_tokens": 183287787.0,
|
| 1432 |
+
"step": 6450
|
| 1433 |
+
},
|
| 1434 |
+
{
|
| 1435 |
+
"entropy": 1.411766264438629,
|
| 1436 |
+
"epoch": 2.2108843537414966,
|
| 1437 |
+
"grad_norm": 1.21875,
|
| 1438 |
+
"learning_rate": 3.929535949452806e-06,
|
| 1439 |
+
"loss": 1.1034652709960937,
|
| 1440 |
+
"mean_token_accuracy": 0.7170097494125366,
|
| 1441 |
+
"num_tokens": 184697327.0,
|
| 1442 |
+
"step": 6500
|
| 1443 |
+
},
|
| 1444 |
+
{
|
| 1445 |
+
"epoch": 2.2108843537414966,
|
| 1446 |
+
"eval_entropy": 1.411220889131562,
|
| 1447 |
+
"eval_loss": 1.1048741340637207,
|
| 1448 |
+
"eval_mean_token_accuracy": 0.7172508204684538,
|
| 1449 |
+
"eval_num_tokens": 184697327.0,
|
| 1450 |
+
"eval_runtime": 8.3646,
|
| 1451 |
+
"eval_samples_per_second": 113.334,
|
| 1452 |
+
"eval_steps_per_second": 14.227,
|
| 1453 |
+
"step": 6500
|
| 1454 |
+
},
|
| 1455 |
+
{
|
| 1456 |
+
"entropy": 1.4167416214942932,
|
| 1457 |
+
"epoch": 2.227891156462585,
|
| 1458 |
+
"grad_norm": 1.0078125,
|
| 1459 |
+
"learning_rate": 3.773503129916374e-06,
|
| 1460 |
+
"loss": 1.1365108489990234,
|
| 1461 |
+
"mean_token_accuracy": 0.7105149948596954,
|
| 1462 |
+
"num_tokens": 186127393.0,
|
| 1463 |
+
"step": 6550
|
| 1464 |
+
},
|
| 1465 |
+
{
|
| 1466 |
+
"entropy": 1.4231563568115235,
|
| 1467 |
+
"epoch": 2.2448979591836733,
|
| 1468 |
+
"grad_norm": 1.046875,
|
| 1469 |
+
"learning_rate": 3.6199077718320817e-06,
|
| 1470 |
+
"loss": 1.146829605102539,
|
| 1471 |
+
"mean_token_accuracy": 0.7080222427845001,
|
| 1472 |
+
"num_tokens": 187593729.0,
|
| 1473 |
+
"step": 6600
|
| 1474 |
+
},
|
| 1475 |
+
{
|
| 1476 |
+
"entropy": 1.4201726150512695,
|
| 1477 |
+
"epoch": 2.261904761904762,
|
| 1478 |
+
"grad_norm": 1.046875,
|
| 1479 |
+
"learning_rate": 3.4688100025512052e-06,
|
| 1480 |
+
"loss": 1.1279085540771485,
|
| 1481 |
+
"mean_token_accuracy": 0.7107222032546997,
|
| 1482 |
+
"num_tokens": 188993851.0,
|
| 1483 |
+
"step": 6650
|
| 1484 |
+
},
|
| 1485 |
+
{
|
| 1486 |
+
"entropy": 1.414864149093628,
|
| 1487 |
+
"epoch": 2.2789115646258504,
|
| 1488 |
+
"grad_norm": 1.0234375,
|
| 1489 |
+
"learning_rate": 3.3202689717040805e-06,
|
| 1490 |
+
"loss": 1.1230371856689454,
|
| 1491 |
+
"mean_token_accuracy": 0.713964912891388,
|
| 1492 |
+
"num_tokens": 190427604.0,
|
| 1493 |
+
"step": 6700
|
| 1494 |
+
},
|
| 1495 |
+
{
|
| 1496 |
+
"entropy": 1.4270901560783387,
|
| 1497 |
+
"epoch": 2.295918367346939,
|
| 1498 |
+
"grad_norm": 1.078125,
|
| 1499 |
+
"learning_rate": 3.1743428280450217e-06,
|
| 1500 |
+
"loss": 1.1262576293945312,
|
| 1501 |
+
"mean_token_accuracy": 0.7121612048149109,
|
| 1502 |
+
"num_tokens": 191836070.0,
|
| 1503 |
+
"step": 6750
|
| 1504 |
+
},
|
| 1505 |
+
{
|
| 1506 |
+
"entropy": 1.4266995477676392,
|
| 1507 |
+
"epoch": 2.312925170068027,
|
| 1508 |
+
"grad_norm": 1.1953125,
|
| 1509 |
+
"learning_rate": 3.0310886966890695e-06,
|
| 1510 |
+
"loss": 1.1296147155761718,
|
| 1511 |
+
"mean_token_accuracy": 0.7117228031158447,
|
| 1512 |
+
"num_tokens": 193246580.0,
|
| 1513 |
+
"step": 6800
|
| 1514 |
+
},
|
| 1515 |
+
{
|
| 1516 |
+
"entropy": 1.4152509379386902,
|
| 1517 |
+
"epoch": 2.3299319727891157,
|
| 1518 |
+
"grad_norm": 1.078125,
|
| 1519 |
+
"learning_rate": 2.890562656749465e-06,
|
| 1520 |
+
"loss": 1.1302790069580078,
|
| 1521 |
+
"mean_token_accuracy": 0.7125261425971985,
|
| 1522 |
+
"num_tokens": 194684126.0,
|
| 1523 |
+
"step": 6850
|
| 1524 |
+
},
|
| 1525 |
+
{
|
| 1526 |
+
"entropy": 1.4296420979499818,
|
| 1527 |
+
"epoch": 2.3469387755102042,
|
| 1528 |
+
"grad_norm": 1.1171875,
|
| 1529 |
+
"learning_rate": 2.752819719384573e-06,
|
| 1530 |
+
"loss": 1.139165802001953,
|
| 1531 |
+
"mean_token_accuracy": 0.7097991979122162,
|
| 1532 |
+
"num_tokens": 196075134.0,
|
| 1533 |
+
"step": 6900
|
| 1534 |
+
},
|
| 1535 |
+
{
|
| 1536 |
+
"entropy": 1.4089247822761535,
|
| 1537 |
+
"epoch": 2.3639455782312924,
|
| 1538 |
+
"grad_norm": 1.0,
|
| 1539 |
+
"learning_rate": 2.6179138062628863e-06,
|
| 1540 |
+
"loss": 1.1174855041503906,
|
| 1541 |
+
"mean_token_accuracy": 0.7136220681667328,
|
| 1542 |
+
"num_tokens": 197502453.0,
|
| 1543 |
+
"step": 6950
|
| 1544 |
+
},
|
| 1545 |
+
{
|
| 1546 |
+
"entropy": 1.4326133465766906,
|
| 1547 |
+
"epoch": 2.380952380952381,
|
| 1548 |
+
"grad_norm": 1.078125,
|
| 1549 |
+
"learning_rate": 2.485897728454546e-06,
|
| 1550 |
+
"loss": 1.137528533935547,
|
| 1551 |
+
"mean_token_accuracy": 0.7098107814788819,
|
| 1552 |
+
"num_tokens": 198927675.0,
|
| 1553 |
+
"step": 7000
|
| 1554 |
+
},
|
| 1555 |
+
{
|
| 1556 |
+
"epoch": 2.380952380952381,
|
| 1557 |
+
"eval_entropy": 1.4109101956631958,
|
| 1558 |
+
"eval_loss": 1.1048496961593628,
|
| 1559 |
+
"eval_mean_token_accuracy": 0.7171741062853517,
|
| 1560 |
+
"eval_num_tokens": 198927675.0,
|
| 1561 |
+
"eval_runtime": 8.35,
|
| 1562 |
+
"eval_samples_per_second": 113.532,
|
| 1563 |
+
"eval_steps_per_second": 14.251,
|
| 1564 |
+
"step": 7000
|
| 1565 |
+
},
|
| 1566 |
+
{
|
| 1567 |
+
"entropy": 1.4286697244644164,
|
| 1568 |
+
"epoch": 2.3979591836734695,
|
| 1569 |
+
"grad_norm": 1.0390625,
|
| 1570 |
+
"learning_rate": 2.3568231657575834e-06,
|
| 1571 |
+
"loss": 1.129428939819336,
|
| 1572 |
+
"mean_token_accuracy": 0.7123888862133027,
|
| 1573 |
+
"num_tokens": 200323695.0,
|
| 1574 |
+
"step": 7050
|
| 1575 |
+
},
|
| 1576 |
+
{
|
| 1577 |
+
"entropy": 1.438240523338318,
|
| 1578 |
+
"epoch": 2.4149659863945576,
|
| 1579 |
+
"grad_norm": 0.94140625,
|
| 1580 |
+
"learning_rate": 2.2307406464670556e-06,
|
| 1581 |
+
"loss": 1.1379161071777344,
|
| 1582 |
+
"mean_token_accuracy": 0.7092709636688233,
|
| 1583 |
+
"num_tokens": 201726356.0,
|
| 1584 |
+
"step": 7100
|
| 1585 |
+
},
|
| 1586 |
+
{
|
| 1587 |
+
"entropy": 1.4395070052146912,
|
| 1588 |
+
"epoch": 2.431972789115646,
|
| 1589 |
+
"grad_norm": 1.0078125,
|
| 1590 |
+
"learning_rate": 2.1076995275949363e-06,
|
| 1591 |
+
"loss": 1.1427032470703125,
|
| 1592 |
+
"mean_token_accuracy": 0.7088804817199708,
|
| 1593 |
+
"num_tokens": 203117543.0,
|
| 1594 |
+
"step": 7150
|
| 1595 |
+
},
|
| 1596 |
+
{
|
| 1597 |
+
"entropy": 1.4223812985420228,
|
| 1598 |
+
"epoch": 2.4489795918367347,
|
| 1599 |
+
"grad_norm": 1.0859375,
|
| 1600 |
+
"learning_rate": 1.987747975548525e-06,
|
| 1601 |
+
"loss": 1.141911849975586,
|
| 1602 |
+
"mean_token_accuracy": 0.7085710716247559,
|
| 1603 |
+
"num_tokens": 204540538.0,
|
| 1604 |
+
"step": 7200
|
| 1605 |
+
},
|
| 1606 |
+
{
|
| 1607 |
+
"entropy": 1.4200545740127564,
|
| 1608 |
+
"epoch": 2.4659863945578233,
|
| 1609 |
+
"grad_norm": 1.1015625,
|
| 1610 |
+
"learning_rate": 1.8709329472749305e-06,
|
| 1611 |
+
"loss": 1.134008255004883,
|
| 1612 |
+
"mean_token_accuracy": 0.7102875554561615,
|
| 1613 |
+
"num_tokens": 205964981.0,
|
| 1614 |
+
"step": 7250
|
| 1615 |
+
},
|
| 1616 |
+
{
|
| 1617 |
+
"entropy": 1.4074384212493896,
|
| 1618 |
+
"epoch": 2.4829931972789114,
|
| 1619 |
+
"grad_norm": 1.0234375,
|
| 1620 |
+
"learning_rate": 1.757300171879045e-06,
|
| 1621 |
+
"loss": 1.1225076293945313,
|
| 1622 |
+
"mean_token_accuracy": 0.7127889108657837,
|
| 1623 |
+
"num_tokens": 207376354.0,
|
| 1624 |
+
"step": 7300
|
| 1625 |
+
},
|
| 1626 |
+
{
|
| 1627 |
+
"entropy": 1.4282367515563965,
|
| 1628 |
+
"epoch": 2.5,
|
| 1629 |
+
"grad_norm": 1.125,
|
| 1630 |
+
"learning_rate": 1.6468941327221377e-06,
|
| 1631 |
+
"loss": 1.1383217620849608,
|
| 1632 |
+
"mean_token_accuracy": 0.7087402617931366,
|
| 1633 |
+
"num_tokens": 208768190.0,
|
| 1634 |
+
"step": 7350
|
| 1635 |
+
},
|
| 1636 |
+
{
|
| 1637 |
+
"entropy": 1.4284853887557984,
|
| 1638 |
+
"epoch": 2.5170068027210886,
|
| 1639 |
+
"grad_norm": 0.98046875,
|
| 1640 |
+
"learning_rate": 1.539758050008141e-06,
|
| 1641 |
+
"loss": 1.1509806060791015,
|
| 1642 |
+
"mean_token_accuracy": 0.7068107867240906,
|
| 1643 |
+
"num_tokens": 210181751.0,
|
| 1644 |
+
"step": 7400
|
| 1645 |
+
},
|
| 1646 |
+
{
|
| 1647 |
+
"entropy": 1.441183557510376,
|
| 1648 |
+
"epoch": 2.534013605442177,
|
| 1649 |
+
"grad_norm": 1.0703125,
|
| 1650 |
+
"learning_rate": 1.4359338638644194e-06,
|
| 1651 |
+
"loss": 1.1412885284423828,
|
| 1652 |
+
"mean_token_accuracy": 0.7094214224815368,
|
| 1653 |
+
"num_tokens": 211592789.0,
|
| 1654 |
+
"step": 7450
|
| 1655 |
+
},
|
| 1656 |
+
{
|
| 1657 |
+
"entropy": 1.4259497380256654,
|
| 1658 |
+
"epoch": 2.5510204081632653,
|
| 1659 |
+
"grad_norm": 1.0703125,
|
| 1660 |
+
"learning_rate": 1.335462217923631e-06,
|
| 1661 |
+
"loss": 1.1345823669433595,
|
| 1662 |
+
"mean_token_accuracy": 0.7104575765132904,
|
| 1663 |
+
"num_tokens": 213028550.0,
|
| 1664 |
+
"step": 7500
|
| 1665 |
+
},
|
| 1666 |
+
{
|
| 1667 |
+
"epoch": 2.5510204081632653,
|
| 1668 |
+
"eval_entropy": 1.4108752523149763,
|
| 1669 |
+
"eval_loss": 1.1047948598861694,
|
| 1670 |
+
"eval_mean_token_accuracy": 0.7172804800402216,
|
| 1671 |
+
"eval_num_tokens": 213028550.0,
|
| 1672 |
+
"eval_runtime": 9.5013,
|
| 1673 |
+
"eval_samples_per_second": 99.775,
|
| 1674 |
+
"eval_steps_per_second": 12.525,
|
| 1675 |
+
"step": 7500
|
| 1676 |
+
},
|
| 1677 |
+
{
|
| 1678 |
+
"entropy": 1.4120974230766297,
|
| 1679 |
+
"epoch": 2.568027210884354,
|
| 1680 |
+
"grad_norm": 1.0625,
|
| 1681 |
+
"learning_rate": 1.2383824434131253e-06,
|
| 1682 |
+
"loss": 1.1174241638183593,
|
| 1683 |
+
"mean_token_accuracy": 0.7148010683059692,
|
| 1684 |
+
"num_tokens": 214461463.0,
|
| 1685 |
+
"step": 7550
|
| 1686 |
+
},
|
| 1687 |
+
{
|
| 1688 |
+
"entropy": 1.42849999666214,
|
| 1689 |
+
"epoch": 2.5850340136054424,
|
| 1690 |
+
"grad_norm": 1.0,
|
| 1691 |
+
"learning_rate": 1.1447325437581214e-06,
|
| 1692 |
+
"loss": 1.1466537475585938,
|
| 1693 |
+
"mean_token_accuracy": 0.7079911732673645,
|
| 1694 |
+
"num_tokens": 215914614.0,
|
| 1695 |
+
"step": 7600
|
| 1696 |
+
},
|
| 1697 |
+
{
|
| 1698 |
+
"entropy": 1.4341383004188537,
|
| 1699 |
+
"epoch": 2.6020408163265305,
|
| 1700 |
+
"grad_norm": 1.0390625,
|
| 1701 |
+
"learning_rate": 1.0545491797046438e-06,
|
| 1702 |
+
"loss": 1.1218763732910155,
|
| 1703 |
+
"mean_token_accuracy": 0.7130703330039978,
|
| 1704 |
+
"num_tokens": 217329434.0,
|
| 1705 |
+
"step": 7650
|
| 1706 |
+
},
|
| 1707 |
+
{
|
| 1708 |
+
"entropy": 1.4117230129241944,
|
| 1709 |
+
"epoch": 2.619047619047619,
|
| 1710 |
+
"grad_norm": 1.234375,
|
| 1711 |
+
"learning_rate": 9.678676549681077e-07,
|
| 1712 |
+
"loss": 1.1172777557373046,
|
| 1713 |
+
"mean_token_accuracy": 0.7140469861030578,
|
| 1714 |
+
"num_tokens": 218728299.0,
|
| 1715 |
+
"step": 7700
|
| 1716 |
+
},
|
| 1717 |
+
{
|
| 1718 |
+
"entropy": 1.4421565866470336,
|
| 1719 |
+
"epoch": 2.6360544217687076,
|
| 1720 |
+
"grad_norm": 1.1484375,
|
| 1721 |
+
"learning_rate": 8.84721902413097e-07,
|
| 1722 |
+
"loss": 1.1367300415039063,
|
| 1723 |
+
"mean_token_accuracy": 0.7104098796844482,
|
| 1724 |
+
"num_tokens": 220137703.0,
|
| 1725 |
+
"step": 7750
|
| 1726 |
+
},
|
| 1727 |
+
{
|
| 1728 |
+
"entropy": 1.4460611391067504,
|
| 1729 |
+
"epoch": 2.6530612244897958,
|
| 1730 |
+
"grad_norm": 0.97265625,
|
| 1731 |
+
"learning_rate": 8.05144470769823e-07,
|
| 1732 |
+
"loss": 1.1389141845703126,
|
| 1733 |
+
"mean_token_accuracy": 0.7093100786209107,
|
| 1734 |
+
"num_tokens": 221566274.0,
|
| 1735 |
+
"step": 7800
|
| 1736 |
+
},
|
| 1737 |
+
{
|
| 1738 |
+
"entropy": 1.433594093322754,
|
| 1739 |
+
"epoch": 2.6700680272108843,
|
| 1740 |
+
"grad_norm": 1.09375,
|
| 1741 |
+
"learning_rate": 7.291665118923919e-07,
|
| 1742 |
+
"loss": 1.1337957000732422,
|
| 1743 |
+
"mean_token_accuracy": 0.7100949645042419,
|
| 1744 |
+
"num_tokens": 223013879.0,
|
| 1745 |
+
"step": 7850
|
| 1746 |
+
},
|
| 1747 |
+
{
|
| 1748 |
+
"entropy": 1.4160945701599121,
|
| 1749 |
+
"epoch": 2.687074829931973,
|
| 1750 |
+
"grad_norm": 1.125,
|
| 1751 |
+
"learning_rate": 6.568177685639187e-07,
|
| 1752 |
+
"loss": 1.1077276611328124,
|
| 1753 |
+
"mean_token_accuracy": 0.7158815062046051,
|
| 1754 |
+
"num_tokens": 224430160.0,
|
| 1755 |
+
"step": 7900
|
| 1756 |
+
},
|
| 1757 |
+
{
|
| 1758 |
+
"entropy": 1.4393118023872375,
|
| 1759 |
+
"epoch": 2.704081632653061,
|
| 1760 |
+
"grad_norm": 1.03125,
|
| 1761 |
+
"learning_rate": 5.881265628532395e-07,
|
| 1762 |
+
"loss": 1.1686244201660156,
|
| 1763 |
+
"mean_token_accuracy": 0.700288165807724,
|
| 1764 |
+
"num_tokens": 225850566.0,
|
| 1765 |
+
"step": 7950
|
| 1766 |
+
},
|
| 1767 |
+
{
|
| 1768 |
+
"entropy": 1.418517336845398,
|
| 1769 |
+
"epoch": 2.7210884353741496,
|
| 1770 |
+
"grad_norm": 1.1171875,
|
| 1771 |
+
"learning_rate": 5.231197850277914e-07,
|
| 1772 |
+
"loss": 1.117198257446289,
|
| 1773 |
+
"mean_token_accuracy": 0.7138521051406861,
|
| 1774 |
+
"num_tokens": 227252331.0,
|
| 1775 |
+
"step": 8000
|
| 1776 |
+
},
|
| 1777 |
+
{
|
| 1778 |
+
"epoch": 2.7210884353741496,
|
| 1779 |
+
"eval_entropy": 1.411157799368145,
|
| 1780 |
+
"eval_loss": 1.1048526763916016,
|
| 1781 |
+
"eval_mean_token_accuracy": 0.7171973621144014,
|
| 1782 |
+
"eval_num_tokens": 227252331.0,
|
| 1783 |
+
"eval_runtime": 9.4712,
|
| 1784 |
+
"eval_samples_per_second": 100.093,
|
| 1785 |
+
"eval_steps_per_second": 12.564,
|
| 1786 |
+
"step": 8000
|
| 1787 |
+
},
|
| 1788 |
+
{
|
| 1789 |
+
"entropy": 1.4183010959625244,
|
| 1790 |
+
"epoch": 2.738095238095238,
|
| 1791 |
+
"grad_norm": 1.0078125,
|
| 1792 |
+
"learning_rate": 4.618228830269822e-07,
|
| 1793 |
+
"loss": 1.1152452850341796,
|
| 1794 |
+
"mean_token_accuracy": 0.7141678440570831,
|
| 1795 |
+
"num_tokens": 228659040.0,
|
| 1796 |
+
"step": 8050
|
| 1797 |
+
},
|
| 1798 |
+
{
|
| 1799 |
+
"entropy": 1.4264278268814088,
|
| 1800 |
+
"epoch": 2.7551020408163263,
|
| 1801 |
+
"grad_norm": 1.0,
|
| 1802 |
+
"learning_rate": 4.0425985250019174e-07,
|
| 1803 |
+
"loss": 1.141219711303711,
|
| 1804 |
+
"mean_token_accuracy": 0.7086251521110535,
|
| 1805 |
+
"num_tokens": 230087322.0,
|
| 1806 |
+
"step": 8100
|
| 1807 |
+
},
|
| 1808 |
+
{
|
| 1809 |
+
"entropy": 1.4321845197677612,
|
| 1810 |
+
"epoch": 2.772108843537415,
|
| 1811 |
+
"grad_norm": 1.078125,
|
| 1812 |
+
"learning_rate": 3.5045322741329435e-07,
|
| 1813 |
+
"loss": 1.1400668334960937,
|
| 1814 |
+
"mean_token_accuracy": 0.7088430786132812,
|
| 1815 |
+
"num_tokens": 231534171.0,
|
| 1816 |
+
"step": 8150
|
| 1817 |
+
},
|
| 1818 |
+
{
|
| 1819 |
+
"entropy": 1.438843650817871,
|
| 1820 |
+
"epoch": 2.7891156462585034,
|
| 1821 |
+
"grad_norm": 0.98046875,
|
| 1822 |
+
"learning_rate": 3.004240712273743e-07,
|
| 1823 |
+
"loss": 1.1484903717041015,
|
| 1824 |
+
"mean_token_accuracy": 0.7076469850540161,
|
| 1825 |
+
"num_tokens": 232975150.0,
|
| 1826 |
+
"step": 8200
|
| 1827 |
+
},
|
| 1828 |
+
{
|
| 1829 |
+
"entropy": 1.425174674987793,
|
| 1830 |
+
"epoch": 2.806122448979592,
|
| 1831 |
+
"grad_norm": 0.9140625,
|
| 1832 |
+
"learning_rate": 2.5419196865309916e-07,
|
| 1833 |
+
"loss": 1.1197071075439453,
|
| 1834 |
+
"mean_token_accuracy": 0.7135571861267089,
|
| 1835 |
+
"num_tokens": 234415037.0,
|
| 1836 |
+
"step": 8250
|
| 1837 |
+
},
|
| 1838 |
+
{
|
| 1839 |
+
"entropy": 1.414315857887268,
|
| 1840 |
+
"epoch": 2.8231292517006805,
|
| 1841 |
+
"grad_norm": 1.1796875,
|
| 1842 |
+
"learning_rate": 2.1177501798397947e-07,
|
| 1843 |
+
"loss": 1.1076421356201172,
|
| 1844 |
+
"mean_token_accuracy": 0.716143217086792,
|
| 1845 |
+
"num_tokens": 235839773.0,
|
| 1846 |
+
"step": 8300
|
| 1847 |
+
},
|
| 1848 |
+
{
|
| 1849 |
+
"entropy": 1.4245373177528382,
|
| 1850 |
+
"epoch": 2.8401360544217686,
|
| 1851 |
+
"grad_norm": 1.125,
|
| 1852 |
+
"learning_rate": 1.7318982401149265e-07,
|
| 1853 |
+
"loss": 1.126745147705078,
|
| 1854 |
+
"mean_token_accuracy": 0.7123495972156525,
|
| 1855 |
+
"num_tokens": 237240977.0,
|
| 1856 |
+
"step": 8350
|
| 1857 |
+
},
|
| 1858 |
+
{
|
| 1859 |
+
"entropy": 1.4268240427970886,
|
| 1860 |
+
"epoch": 2.857142857142857,
|
| 1861 |
+
"grad_norm": 0.9375,
|
| 1862 |
+
"learning_rate": 1.3845149152487803e-07,
|
| 1863 |
+
"loss": 1.1193339538574218,
|
| 1864 |
+
"mean_token_accuracy": 0.7143187963962555,
|
| 1865 |
+
"num_tokens": 238665599.0,
|
| 1866 |
+
"step": 8400
|
| 1867 |
+
},
|
| 1868 |
+
{
|
| 1869 |
+
"entropy": 1.4306807327270508,
|
| 1870 |
+
"epoch": 2.8741496598639458,
|
| 1871 |
+
"grad_norm": 1.140625,
|
| 1872 |
+
"learning_rate": 1.0757361939813137e-07,
|
| 1873 |
+
"loss": 1.12787353515625,
|
| 1874 |
+
"mean_token_accuracy": 0.7113662779331207,
|
| 1875 |
+
"num_tokens": 240099917.0,
|
| 1876 |
+
"step": 8450
|
| 1877 |
+
},
|
| 1878 |
+
{
|
| 1879 |
+
"entropy": 1.4395443654060365,
|
| 1880 |
+
"epoch": 2.891156462585034,
|
| 1881 |
+
"grad_norm": 1.171875,
|
| 1882 |
+
"learning_rate": 8.056829526649767e-08,
|
| 1883 |
+
"loss": 1.1395121002197266,
|
| 1884 |
+
"mean_token_accuracy": 0.7095162570476532,
|
| 1885 |
+
"num_tokens": 241504767.0,
|
| 1886 |
+
"step": 8500
|
| 1887 |
+
},
|
| 1888 |
+
{
|
| 1889 |
+
"epoch": 2.891156462585034,
|
| 1890 |
+
"eval_entropy": 1.4105532329623438,
|
| 1891 |
+
"eval_loss": 1.1048182249069214,
|
| 1892 |
+
"eval_mean_token_accuracy": 0.7172953571592059,
|
| 1893 |
+
"eval_num_tokens": 241504767.0,
|
| 1894 |
+
"eval_runtime": 9.5391,
|
| 1895 |
+
"eval_samples_per_second": 99.381,
|
| 1896 |
+
"eval_steps_per_second": 12.475,
|
| 1897 |
+
"step": 8500
|
| 1898 |
+
},
|
| 1899 |
+
{
|
| 1900 |
+
"entropy": 1.4125055813789367,
|
| 1901 |
+
"epoch": 2.9081632653061225,
|
| 1902 |
+
"grad_norm": 1.0078125,
|
| 1903 |
+
"learning_rate": 5.7446090794580636e-08,
|
| 1904 |
+
"loss": 1.1209735870361328,
|
| 1905 |
+
"mean_token_accuracy": 0.713252090215683,
|
| 1906 |
+
"num_tokens": 242946321.0,
|
| 1907 |
+
"step": 8550
|
| 1908 |
+
},
|
| 1909 |
+
{
|
| 1910 |
+
"entropy": 1.4286201095581055,
|
| 1911 |
+
"epoch": 2.925170068027211,
|
| 1912 |
+
"grad_norm": 1.0859375,
|
| 1913 |
+
"learning_rate": 3.821605753789648e-08,
|
| 1914 |
+
"loss": 1.1291080474853517,
|
| 1915 |
+
"mean_token_accuracy": 0.712302508354187,
|
| 1916 |
+
"num_tokens": 244378684.0,
|
| 1917 |
+
"step": 8600
|
| 1918 |
+
},
|
| 1919 |
+
{
|
| 1920 |
+
"entropy": 1.4226014256477355,
|
| 1921 |
+
"epoch": 2.942176870748299,
|
| 1922 |
+
"grad_norm": 1.015625,
|
| 1923 |
+
"learning_rate": 2.288572339949835e-08,
|
| 1924 |
+
"loss": 1.1374082946777344,
|
| 1925 |
+
"mean_token_accuracy": 0.7114299070835114,
|
| 1926 |
+
"num_tokens": 245806693.0,
|
| 1927 |
+
"step": 8650
|
| 1928 |
+
},
|
| 1929 |
+
{
|
| 1930 |
+
"entropy": 1.4223952269554139,
|
| 1931 |
+
"epoch": 2.9591836734693877,
|
| 1932 |
+
"grad_norm": 1.0859375,
|
| 1933 |
+
"learning_rate": 1.1461089683053595e-08,
|
| 1934 |
+
"loss": 1.1320287322998046,
|
| 1935 |
+
"mean_token_accuracy": 0.7105390858650208,
|
| 1936 |
+
"num_tokens": 247211598.0,
|
| 1937 |
+
"step": 8700
|
| 1938 |
+
},
|
| 1939 |
+
{
|
| 1940 |
+
"entropy": 1.4174312782287597,
|
| 1941 |
+
"epoch": 2.9761904761904763,
|
| 1942 |
+
"grad_norm": 1.0234375,
|
| 1943 |
+
"learning_rate": 3.946628743543013e-09,
|
| 1944 |
+
"loss": 1.1252450561523437,
|
| 1945 |
+
"mean_token_accuracy": 0.713299845457077,
|
| 1946 |
+
"num_tokens": 248623558.0,
|
| 1947 |
+
"step": 8750
|
| 1948 |
+
},
|
| 1949 |
+
{
|
| 1950 |
+
"entropy": 1.430122504234314,
|
| 1951 |
+
"epoch": 2.9931972789115644,
|
| 1952 |
+
"grad_norm": 1.1640625,
|
| 1953 |
+
"learning_rate": 3.452822364835573e-10,
|
| 1954 |
+
"loss": 1.142134246826172,
|
| 1955 |
+
"mean_token_accuracy": 0.7098831164836884,
|
| 1956 |
+
"num_tokens": 250064804.0,
|
| 1957 |
+
"step": 8800
|
| 1958 |
+
},
|
| 1959 |
+
{
|
| 1960 |
+
"entropy": 1.4038671672344207,
|
| 1961 |
+
"epoch": 3.0,
|
| 1962 |
+
"mean_token_accuracy": 0.7150700688362122,
|
| 1963 |
+
"num_tokens": 250633827.0,
|
| 1964 |
+
"step": 8820,
|
| 1965 |
+
"total_flos": 1.3775254383370368e+18,
|
| 1966 |
+
"train_loss": 1.158577897164827,
|
| 1967 |
+
"train_runtime": 2757.2959,
|
| 1968 |
+
"train_samples_per_second": 102.347,
|
| 1969 |
+
"train_steps_per_second": 3.199
|
| 1970 |
+
}
|
| 1971 |
+
],
|
| 1972 |
+
"logging_steps": 50,
|
| 1973 |
+
"max_steps": 8820,
|
| 1974 |
+
"num_input_tokens_seen": 0,
|
| 1975 |
+
"num_train_epochs": 3,
|
| 1976 |
+
"save_steps": 500,
|
| 1977 |
+
"stateful_callbacks": {
|
| 1978 |
+
"TrainerControl": {
|
| 1979 |
+
"args": {
|
| 1980 |
+
"should_epoch_stop": false,
|
| 1981 |
+
"should_evaluate": false,
|
| 1982 |
+
"should_log": false,
|
| 1983 |
+
"should_save": true,
|
| 1984 |
+
"should_training_stop": true
|
| 1985 |
+
},
|
| 1986 |
+
"attributes": {}
|
| 1987 |
+
}
|
| 1988 |
+
},
|
| 1989 |
+
"total_flos": 1.3775254383370368e+18,
|
| 1990 |
+
"train_batch_size": 32,
|
| 1991 |
+
"trial_name": null,
|
| 1992 |
+
"trial_params": null
|
| 1993 |
+
}
|