Commit History

Add qa style data for alpaca instructions, fix one_cycle scheduler
3a50377

winglian commited on

Merge pull request #34 from OpenAccess-AI-Collective/dev-unstable
b029a11
unverified

winglian commited on

cuda/pytorch matrix builds
e3df3a9

winglian commited on

cuda, pytorch matrix for base builds
f950a88

winglian commited on

don't need to set here
de6da13

winglian commited on

be able to use adam bnb 8bit and one cycle scheduler w fsdp
9493b1b

winglian commited on

Update src/axolotl/utils/models.py for info msg
1b3e401
unverified

winglian Nanobit commited on

Update scripts/finetune.py
3457810
unverified

winglian Nanobit commited on

Update scripts/finetune.py for logging
ae1719d
unverified

winglian Nanobit commited on

Update src/axolotl/utils/data.py for spelling
98a6781
unverified

winglian Nanobit commited on

make sure to use train split if loading from hf
607a4d3

winglian commited on

make one cycle lr div factor configurable
99383f1

winglian commited on

fix new dataset prompt tokenizers
0f74464

winglian commited on

add missing __init__
e0602a9

winglian commited on

pygmalion dataset prompts format, cached tokenized datasets should be hashed on the tokenizer too
2809f3f

winglian commited on

tokenization fixes
4ea9a66

winglian commited on

Merge pull request #32 from NanoCode012/patch-2
ed37b22
unverified

winglian commited on

optionally be able to specify alpaca or chat style prompts
1d5ab84

winglian commited on

Set `half` using `cfg.fp16` for 4bit
641f801
unverified

Nanobit commited on

update entrypoint and force min accelerate
fa8bd14

winglian commited on

concise multiple choice and tldr summarize
1365073

winglian commited on

support for replit lm
8c2f3cb

winglian commited on

add alpaca multiple choice instruct dataset support
b46bc02

winglian commited on

Merge pull request #29 from NanoCode012/patch-1
e553c90
unverified

winglian commited on

Add `lora_modules_to_save`
2c73c81
unverified

Nanobit commited on

reorder options so debug can happen in the same prepare step
f98e173

winglian commited on

fix prompters, especially the sharegpt prompter
5e37144

winglian commited on

more fixes
bdbca8f

winglian commited on

more fixes
42410c7

winglian commited on

fix torch_dtype for model load
aef00b6

winglian commited on

move filter to before saving so it doesn't happen everytime, update runpod manual script
0d28df0

winglian commited on

whoops, gt vs lt
84c7bc4

winglian commited on

optimize dataloading to use cache, fix model token embedding sizes
aa3c3f9

winglian commited on

Merge pull request #25 from NanoCode012/patch-2
f6d1fa4
unverified

winglian commited on

Merge branch 'main' into patch-2
89b7f26
unverified

Nanobit commited on

fix config for parity with previous change
165da58

winglian commited on

Merge pull request #27 from NanoCode012/patch-1
4cc7ed8
unverified

winglian commited on

Fix typo
52aada7
unverified

Nanobit commited on

Merge pull request #26 from OpenAccess-AI-Collective/mpt-triton
688c73a
unverified

winglian commited on

black formatting
2bc1a5b

winglian commited on

various fixes
7a490a4

winglian commited on

Fix Trainer() got multiple values for keyword argument 'callbacks'
813aab3
unverified

Nanobit commited on

testing mpt triton
e2e68c3

winglian commited on

fix conditional so alpaca doesn't choke
a27d594

winglian commited on

Merge pull request #23 from NanoCode012/patch-1
1fb0376
unverified

winglian commited on

Update finetune.py
915c56c
unverified

winglian commited on

not everyone has bf16 available
df9c508

winglian commited on

add 4bit lora 7b
7967cd1

winglian commited on

Don't save full model for lora
cd23959
unverified

Nanobit commited on

Save adapter for lora
71a1f7f
unverified

Nanobit commited on