tocmo0nlord
0 Followers · 0 Following
  • Joined on 2026-02-11
Repositories
11
 Projects Packages Public Activity Starred Repositories
tocmo0nlord created branch no-zero-ds-train in tocmo0nlord/axolotl 2026-04-26 15:07:40 +00:00
tocmo0nlord pushed to no-zero-ds-train at tocmo0nlord/axolotl 2026-04-26 15:07:40 +00:00
9f1d548534 don't use zero first context for loading datasets
tocmo0nlord created branch offload-activations-disk in tocmo0nlord/axolotl 2026-04-26 15:07:40 +00:00
tocmo0nlord pushed to offload-activations-disk at tocmo0nlord/axolotl 2026-04-26 15:07:40 +00:00
6100baea0d offload activations to disk instead of CPU RAM
tocmo0nlord created branch olmo-no-position_ids in tocmo0nlord/axolotl 2026-04-26 15:07:40 +00:00
tocmo0nlord pushed to olmo-no-position_ids at tocmo0nlord/axolotl 2026-04-26 15:07:40 +00:00
7c5aa4791f drop position_ids for olmo model
tocmo0nlord created branch online-topk-kd in tocmo0nlord/axolotl 2026-04-26 15:07:40 +00:00
tocmo0nlord pushed to online-topk-kd at tocmo0nlord/axolotl 2026-04-26 15:07:40 +00:00
b8d52a2193 use kwargs
002b1ac967 max new tokens for online generation
17b01bfe36 handle input only for online
a0669335e2 online top-k kd
Compare 4 commits »
tocmo0nlord created branch openorca in tocmo0nlord/axolotl 2026-04-26 15:07:40 +00:00
tocmo0nlord pushed to neft-v2 at tocmo0nlord/axolotl 2026-04-26 15:07:39 +00:00
080612219b use even if not using sample packing
f95858d369 alternate impl of NEFT
Compare 2 commits »
tocmo0nlord created branch no-bos-tokens-packing in tocmo0nlord/axolotl 2026-04-26 15:07:39 +00:00
tocmo0nlord pushed to no-bos-tokens-packing at tocmo0nlord/axolotl 2026-04-26 15:07:39 +00:00
05d19d2037 remove debugging, use gpt2 since starcoder requires consent
61f44f311e fix packing for tokenizers that don't use a bos_token when the bos token and eos token are both the same
Compare 2 commits »
tocmo0nlord created branch no-seq-len in tocmo0nlord/axolotl 2026-04-26 15:07:39 +00:00
tocmo0nlord pushed to no-seq-len at tocmo0nlord/axolotl 2026-04-26 15:07:39 +00:00
c3db6dd307 remove hardcode
9a6e9d8d15 no sequence length support
Compare 2 commits »
tocmo0nlord pushed to muon-validation at tocmo0nlord/axolotl 2026-04-26 15:07:38 +00:00
c578c8f256 Validation for Muon optimizer with DS/FSDP
tocmo0nlord created branch nca-pair in tocmo0nlord/axolotl 2026-04-26 15:07:38 +00:00
tocmo0nlord pushed to nca-pair at tocmo0nlord/axolotl 2026-04-26 15:07:38 +00:00
317761406e add support for NCA
tocmo0nlord created branch nd_parallel in tocmo0nlord/axolotl 2026-04-26 15:07:38 +00:00
tocmo0nlord pushed to nd_parallel at tocmo0nlord/axolotl 2026-04-26 15:07:38 +00:00
bc2bc688d8 update fsdp2 patch
b3c04dd9fe workaround for fsdp2 optimizer save failures
972c719d38 use latest transformers on main with fix
2c1cb8b300 fix for accelerator state getting reset and missing schema
cca207eec4 handle none checks
Compare 9 commits »
tocmo0nlord created branch neft-v2 in tocmo0nlord/axolotl 2026-04-26 15:07:38 +00:00