tocmo0nlord/axolotl
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

feat: Add Mistral Medium 3.5 (#3633)

Browse Source 
* fix: clarify incompat

* fix: transformers api change upstream

* fix: add pre prop

* feat: add examples

* chore: cleanup

* chore: update readme
This commit is contained in:
NanoCode012
2026-04-29 22:46:51 +07:00
committed by GitHub
parent ac77da96da
commit ebbd7fa847
9 changed files with 210 additions and 7 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
examples/ministral3/ministral3-3b-qlora.yaml
View File

@@ -59,7 +59,7 @@ gradient_checkpointing: true
resume_from_checkpoint:
logging_steps: 1
flash_attention: true
scaling_softmax: true
# scaling_softmax: true # needs flex_attention
warmup_ratio: 0.1
evals_per_epoch: 1