prithivMLmods/SmolLM2_135M_Grpo_Checkpoint
Partial fine-tune vs HuggingFaceTB/SmolLM2-135M-Instruct. Partial change: 21% of the weights changed (embeddings, norms), in 29 of 30 layers.
Tensors changed: Embeddings 1/1, Norms 47/61, MLP 0/90, Attention 0/120
2 derivatives checked
- mradermacher/SmolLM2_135M_Grpo_Checkpoint-i1-GGUF: Quantization
- mradermacher/SmolLM2_135M_Grpo_Checkpoint-GGUF: Same weights, converted
2 other repos hold byte-for-byte identical weight files.