Spaces:

Dovakiins
/

qwerrwe

Build error

App Files Files Community

qwerrwe / src /axolotl /utils

Ctrl+K

Ctrl+K

100 contributors

History: 254 commits

chargoddard's picture

ReLoRA implementation (with quantization) (#322)

bde3c5a unverified almost 2 years ago

__init__.py

0 Bytes

WIP large refactor to make finetune script a little more manageable (#3) over 2 years ago
bench.py

1.24 kB

Fix crash when running without CUDA almost 2 years ago
callbacks.py

2.81 kB

ReLoRA implementation (with quantization) (#322) almost 2 years ago
collators.py

5.39 kB

Attention mask and position id fixes for packing (#285) almost 2 years ago
config.py

8.19 kB

ReLoRA implementation (with quantization) (#322) almost 2 years ago
data.py
24 kB

support user defined prompters, pretokenized datasets in config, local parquet, local arrow files (#348) almost 2 years ago
dataloader.py

9.86 kB

workaround so training doesn't hang when packed dataloader batches aren't even (#461) almost 2 years ago
dict.py

326 Bytes

fix DefaultDict.__or__ almost 2 years ago
distributed.py

1.31 kB

use context manager to run things on rank0 before others (#397) almost 2 years ago
models.py
19.4 kB

recast loralayer, norm, lmhead + embed token weights per original qlora (#393) almost 2 years ago
schedulers.py

3.58 kB

add axolotl trainer and quadratic warmup about 2 years ago
tokenization.py

1.37 kB

Adding logging enhancement about 2 years ago
trainer.py
25.7 kB

ReLoRA implementation (with quantization) (#322) almost 2 years ago
wandb.py

858 Bytes

Add wandb_entity to wandb options, update example configs, update README (#361) almost 2 years ago