|
--- |
|
library_name: transformers |
|
license: llama3.1 |
|
base_model: meta-llama/Llama-3.1-8B |
|
tags: |
|
- axolotl |
|
- generated_from_trainer |
|
model-index: |
|
- name: Llama3.1-8B-v0.1-dolma-skymizer-method-0.6 |
|
results: [] |
|
--- |
|
|
|
<!-- This model card has been generated automatically according to the information the Trainer had access to. You |
|
should probably proofread and complete it, then remove this comment. --> |
|
|
|
[<img src="https://raw.githubusercontent.com/axolotl-ai-cloud/axolotl/main/image/axolotl-badge-web.png" alt="Built with Axolotl" width="200" height="32"/>](https://github.com/axolotl-ai-cloud/axolotl) |
|
<details><summary>See axolotl config</summary> |
|
|
|
axolotl version: `0.5.2` |
|
```yaml |
|
base_model: meta-llama/Llama-3.1-8B |
|
model_type: AutoModelForCausalLM |
|
tokenizer_type: AutoTokenizer |
|
tokenizer_use_fast: false |
|
resize_token_embeddings_to_32x: false |
|
|
|
flash_attention: true |
|
xformers_attention: |
|
|
|
load_in_8bit: false |
|
load_in_4bit: false |
|
strict: false |
|
|
|
datasets: |
|
- path: skymizer/Llama3.1-base-tokenized-dolma-v1_7-50B |
|
train_on_split: train |
|
type: completion |
|
|
|
test_datasets: |
|
- path: skymizer/Llama3.1-tokenized-dolma-v1_7-test |
|
split: test |
|
type: completion |
|
|
|
is_preprocess: true |
|
skip_prepare_dataset: true |
|
|
|
dataset_prepared_path: /mnt/home/model-team/datasets/pretokenized/Llama3.1-8B-base-tokenized-dolma-v1_7_50B-4096 |
|
|
|
hf_use_auth_token: true |
|
output_dir: /mnt/home/model-team/models/Llama3.1-8B-v0.1-STE-0.6 |
|
resume_from_checkpoint: |
|
auto_resume_from_checkpoints: true |
|
|
|
sequence_len: 4096 |
|
sample_packing: true |
|
sample_packing_group_size: 100000 |
|
sample_packing_bin_size: 200 |
|
pad_to_sequence_len: true |
|
|
|
eval_sample_packing: false |
|
# eval_causal_lm_metrics: ["perplexity"] |
|
|
|
wandb_project: "sparse-tuning-cpt" |
|
wandb_entity: |
|
wandb_watch: |
|
wandb_name: "Llama3.1-8B-v0.1-dolma-STE-0.6" |
|
wandb_log_model: |
|
|
|
# global batch size = 2 * 8 * 8 GPUs * 8 Nodes * 4096 = 4M |
|
gradient_accumulation_steps: 8 |
|
micro_batch_size: 2 |
|
eval_batch_size: 1 |
|
max_steps: 10000 |
|
optimizer: adamw_torch |
|
learning_rate: 0.00005 |
|
lr_scheduler: cosine |
|
cosine_min_lr_ratio: 0.2 |
|
weight_decay: 0.0 |
|
adam_beta1: 0.9 |
|
adam_beta2: 0.95 |
|
adam_eps: 0.000001 |
|
max_grad_norm: 1.0 |
|
|
|
train_on_inputs: false |
|
group_by_length: false |
|
bf16: true |
|
fp16: |
|
tf32: false |
|
|
|
hub_model_id: "skymizer/Llama3.1-8B-v0.1-dolma-skymizer-method-0.6" |
|
|
|
save_strategy: "steps" |
|
save_steps: 500 |
|
|
|
gradient_checkpointing: true |
|
gradient_checkpointing_kwargs: |
|
use_reentrant: false |
|
early_stopping_patience: |
|
resume_from_checkpoint: |
|
local_rank: |
|
logging_steps: 1 |
|
|
|
warmup_steps: 375 |
|
eval_steps: 500 |
|
eval_table_size: |
|
debug: |
|
deepspeed: /root/train/axolotl/deepspeed_configs/zero3_bf16.json |
|
fsdp: |
|
fsdp_config: |
|
seed: 42 |
|
|
|
special_tokens: |
|
pad_token: "<|end_of_text|>" |
|
|
|
``` |
|
|
|
</details><br> |
|
|
|
# Llama3.1-8B-v0.1-dolma-skymizer-method-0.6 |
|
|
|
This model is a fine-tuned version of [meta-llama/Llama-3.1-8B](https://huggingface.co/meta-llama/Llama-3.1-8B) on an unknown dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 2.3883 |
|
|
|
## Model description |
|
|
|
More information needed |
|
|
|
## Intended uses & limitations |
|
|
|
More information needed |
|
|
|
## Training and evaluation data |
|
|
|
More information needed |
|
|
|
## Training procedure |
|
|
|
### Training hyperparameters |
|
|
|
The following hyperparameters were used during training: |
|
- learning_rate: 5e-05 |
|
- train_batch_size: 2 |
|
- eval_batch_size: 1 |
|
- seed: 42 |
|
- distributed_type: multi-GPU |
|
- num_devices: 64 |
|
- gradient_accumulation_steps: 8 |
|
- total_train_batch_size: 1024 |
|
- total_eval_batch_size: 64 |
|
- optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.95) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments |
|
- lr_scheduler_type: cosine |
|
- lr_scheduler_warmup_steps: 375 |
|
- training_steps: 10000 |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | |
|
|:-------------:|:------:|:-----:|:---------------:| |
|
| 2.2837 | 0.0001 | 1 | 2.5425 | |
|
| 2.2557 | 0.0414 | 500 | 2.4568 | |
|
| 2.2641 | 0.0829 | 1000 | 2.4520 | |
|
| 2.2207 | 0.1243 | 1500 | 2.4477 | |
|
| 2.3003 | 0.1657 | 2000 | 2.4432 | |
|
| 2.2382 | 0.2072 | 2500 | 2.4388 | |
|
| 2.2339 | 0.2486 | 3000 | 2.4349 | |
|
| 2.2517 | 0.2901 | 3500 | 2.4303 | |
|
| 2.2483 | 0.3315 | 4000 | 2.4246 | |
|
| 2.2067 | 0.3729 | 4500 | 2.4207 | |
|
| 2.2485 | 0.4144 | 5000 | 2.4163 | |
|
| 2.2541 | 0.4558 | 5500 | 2.4123 | |
|
| 2.2192 | 0.4972 | 6000 | 2.4084 | |
|
| 2.2346 | 0.5387 | 6500 | 2.4041 | |
|
| 2.2106 | 0.5801 | 7000 | 2.4010 | |
|
| 2.2112 | 0.6215 | 7500 | 2.3982 | |
|
| 2.2215 | 0.6630 | 8000 | 2.3951 | |
|
| 2.2118 | 0.7044 | 8500 | 2.3924 | |
|
| 2.1933 | 0.7458 | 9000 | 2.3905 | |
|
| 2.1813 | 0.7873 | 9500 | 2.3893 | |
|
| 2.1969 | 0.8287 | 10000 | 2.3883 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.46.3 |
|
- Pytorch 2.5.1+cu124 |
|
- Datasets 3.1.0 |
|
- Tokenizers 0.20.3 |
|
|
