arvindanand/ValidateAI-3-33B-Ties
Full fine-tune vs deepseek-ai/deepseek-coder-33b-instruct. Fine-tune: 100% of the weights changed, across 62 of 62 layers.
Tensors changed: Embeddings 1/1, Norms 125/125, MLP 186/186, Attention 248/248, lm_head 1/1
1 derivatives checked
- mradermacher/ValidateAI-3-33B-Ties-GGUF: Quantization
1 other repos hold byte-for-byte identical weight files.