Naphon/pythia-2.8b-thai-base-v1
LoRA-style vs EleutherAI/pythia-2.8b-deduped. Looks like a merged LoRA (or a fine-tune with frozen norms): only the weight matrices of dense, dense_4h_to_h, dense_h_to_4h, query_key_value changed, in 32 of 32 layers; norms, biases and embeddings are untouched.
Tensors changed: Norms 0/130, Attention 64/128, MLP 64/128
1 derivatives checked
- mradermacher/pythia-2.8b-thai-base-v1-GGUF: Same weights, converted