Commit History
feat: remove need to add load_in* during merge (#1017)
f6ecf14
unverified
Nanobit
commited on
remove landmark attn and xpos rope implementations (#1010)
70b46ca
unverified
winglian
commited on
add config to model card (#1005)
85dd4d5
unverified
hamel
commited on
FEAT: add tagging support to axolotl (#1004)
db9094d
unverified
Feat: Warns to add to modules_to_save when adding tokens or switching special_tokens (#787)
1ffa386
unverified
Nanobit
commited on
fix mistral prompt assembly (#982)
7bbaac9
unverified
hamel
commited on
fix: add lr scheduler kwargs to Trainer (#972)
13e9381
unverified
Nanobit
commited on
Fix prompt assembly for llama (#952)
5ada140
unverified
fix: switch to using the HuggingFace Transformers NEFT implementation (#941)
ef24342
unverified
kallewoof
commited on
Fix Deepspeed loading (#950)
5ea3aa3
unverified
winglian
commited on
Flash attn hotfix (#951)
f1f60cb
unverified
winglian
commited on
fix: remove excessive newlines in system prompt(s) for alpaca (#936)
450e04d
unverified
kallewoof
commited on
new evals_per_epoch and saves_per_epoch to make things cleaner (#944)
5f79b82
unverified
winglian
commited on
Respect sequence_len in config for `type: llama2_chat` (#926)
f1de29d
unverified
hamel
commited on
Mixtral official (#942)
7fabc4d
unverified
winglian
commited on
Mixtral: More correct MoE, lower loss (#932)
86487c2
unverified
casperhansen
commited on
Mixtral multipack (#928)
68b227a
unverified
winglian
commited on
fixing prompt template of chatml by removal of linebreak (#922)
03c6318
unverified
support for mamba (#915)
40a6362
unverified
winglian
commited on
fix(tokenizer): handle fast tokenizer properly for bos/eos (#914)
fde091c
unverified
Nanobit
commited on
Support device_map=sequential & max_memory config parameters (#903)
992e742
unverified
Feat(wandb): Refactor to be more flexible (#767)
a1da39c
unverified
Nanobit
commited on
feature: loss watchdog for terminating training runs that are failing (#899)
58ec8b1
unverified
fix for qwen w lora (#906)
3e3229e
unverified
winglian
commited on
ensure merged model matches the training dtype (#902)
1d21aa6
unverified
winglian
commited on
Feat: Add Qwen (#894)
1115c50
unverified
Nanobit
commited on
fix: warning should not show if eval_batch_size not provided (#896)
7ee3c4c
unverified
Nanobit
commited on
Feat: Add warmup_ratio (#893)
fb12895
unverified
Nanobit
commited on
fix: revert local dir dataset load (#878)
575a082
unverified
Nanobit
commited on
Phi update 202311 (#876)
9bf854e
unverified
winglian
commited on
don't train if eval split is too small (#873)
797f3dd
unverified
winglian
commited on
Feat: Add dataset loading from S3, GCS (#765)
3cc67d2
unverified
Nanobit
commited on
allow overriding of model_config parameters from the YML (#853)
1bc1186
unverified
winglian
commited on
multipack len should use max, not min (#863)
0c2a630
unverified
winglian
commited on
adds llama and mistral dropout support (#858)
db8a8af
unverified
winglian
commited on
various bugfixes (#856)
1470650
unverified
winglian
commited on
include the suffix modified string in ascii art (#852)
614cff4
unverified
fpreiss
commited on
cleanup the old multipack dataloader (#841)
1a6309c
unverified
winglian
commited on
multipack w batch sampler (#795)
641e6f7
unverified
winglian
commited on
use accelerate logging for zero/main loggin only
b2430ce
winglian
commited on
cleanup verbosity a bit
4c834bf
winglian
commited on
Feat: Added Gradio support (#812)
738a057
unverified
stillerman
commited on
update table for rwkv4 support, fix process count for dataset (#822)
cdc71f7
unverified
winglian
commited on
fix model parallel (#816)
964d858
unverified
winglian
commited on
fix(tokenizer): update log order after update (#806)
10388a8
unverified
Nanobit
commited on
fix(config): Set eos/bos to tokenizer if different (#801)
637ed09
unverified
Nanobit
commited on