Hugging Face's logo

natmin322
/
Continual

Model card Files
xet
 Community
Continual / improve_gainlora /src

Commit History

rls t5 large
9f858c3

natmin322 commited on

rls t5 large
2c4cffd

natmin322 commited on

v9-fix: oracle training routing, calibrated inference, update docs
5a14212

natmin322 commited on

v7: C5 Data-Informed Subspace Init + restructure contributions to 2 core claims
aeb2d78

natmin322 commited on

fix: reduce CPU RAM to prevent OOM SIGKILL - gc.collect, del temps, eval_accumulation_steps, cache cleanup
c03ffe2

natmin322 commited on

fix: add trust_remote_code=True to second load_dataset call in run_llama.py
e9acf77

natmin322 commited on

dataset: allow custom dataset code by setting trust_remote_code=True in run_llama.py
55f7d25

natmin322 commited on

C4: Spectrally-Conditioned LoRA Training — preconditioned gradient + spectral entropy regularization
2d42b51

natmin322 commited on

SpecRoute V3: adaptive bias, symmetric inference, threshold 0.995, batch size optimization
9ea634d

natmin322 commited on

fix llama
99e2af7

natmin322 commited on

improve v2
aca5a60

natmin322 commited on

fix: pass attention_mask directly to model.generate(), not via GenerationConfig
915a112

natmin322 commited on

fix: override _save to disable safetensors for T5 shared embedding weights
bb4c9d9

natmin322 commited on

fix: denumpify_detensorize moved to trainer_utils in transformers 4.40+
a57a027

natmin322 commited on

fix: add explicit trainer_pt_utils imports (nested_truncate etc.) removed from trainer.* in 4.40+
e4e078c

natmin322 commited on

fix: comprehensive transformers 4.40 compat across all trainer files
2e720a9

natmin322 commited on

fix: replace removed _pad_across_processes with accelerator API
164f658

natmin322 commited on