Spaces:

Dovakiins
/

qwerrwe

Build error

App Files Files Community

qwerrwe / src /axolotl /monkeypatch

Ctrl+K

Ctrl+K

100 contributors

History: 21 commits

tmm1's picture

update flash-attn patch for 70B/GQA and inference using helper from flash-attn tests

d773384 almost 2 years ago

llama_attn_hijack_flash.py
11.2 kB

update flash-attn patch for 70B/GQA and inference using helper from flash-attn tests almost 2 years ago
llama_attn_hijack_sdp.py

4.67 kB

split sdp attn into its own patch almost 2 years ago
llama_attn_hijack_xformers.py

5.55 kB

sync xformers patch to follow shared format and be diffable almost 2 years ago
llama_expand_mask.py

1.92 kB

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

47.8 kB

Fixed pre-commit problems, fixed small bug in logging_config to handle LOG_LEVEL env var almost 2 years ago
utils.py

3.87 kB

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

3.34 kB

add support to extend context with xpos rope almost 2 years ago