use fastchat conversations template (#578)

update for recent transformers updates (#636)

Feat: Add support for upstream FA2 (#626)

btlm and falcon monkey patches for flash attn (#566)

Add training callback to send predictions to WandB table (#521)

reorg a bit

use flash_attn rmsnorm when available (#526)

use flash_attn xentropy when available (#525)

fix checkpints on multigpu (#481)

ReLoRA implementation (with quantization) (#322)

fix eval regression caused in 13f7efaf74fcd3c4514277ccb71914c589873f6a

is_causal fix for evals?

fix evals (#447)

standardize attn hijack patches (#381)

fix check for flash attn branching (#377)

Attention mask and position id fixes for packing (#285)

Update XFormers Attention Monkeypatch to handle Llama-2 70B (GQA) (#339)

move flash-attn monkey patch alongside the others

fix sdp attention to use the flash/mem-efficient context manaager

Fixed pre-commit problems, fixed small bug in logging_config to handle LOG_LEVEL env var

Adding logging enhancement

Fix set mem_id for inference and refactor

Clean up landmark patching

Refactor landmark attention patch

add support to extend context with xpos rope

Fix grad checkpoint and outputs param

Feat: Add landmark attention

don't worry about dupes

Update src/axolotl/monkeypatch/llama_attn_hijack_xformers.py

Update src/axolotl/monkeypatch/llama_attn_hijack_xformers.py

black formatting

copy xformers attn from ooba since we removed dep on alpaca_lora_4bit