Pythia-410M Code Domain Specialist

Fine-tuned from EleutherAI/pythia-410m-deduped on code domain data as part of the Symbiogenesis model merging experiments.

Training Details

  • Base: EleutherAI/pythia-410m-deduped (410M params)
  • Domain: code
  • Steps: 500
  • Hardware: Apple M3 Ultra (MPS)
  • This is a pre-fusion checkpoint used as input to DARE, TIES, and symbiogenesis merging methods.

Links

Downloads last month
11
Safetensors
Model size
0.4B params
Tensor type
F32
·
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for LisaMegaWatts/pythia-410m-code-specialist

Finetuned
(27)
this model