Jainamshahhh/helalm-4b
LoRA-style (fewer layers) vs google/gemma-3-4b-it. Jainamshahhh/helalm-4b matches part of google/gemma-3-4b-it (444 of its 883 tensors; the rest of google/gemma-3-4b-it has no counterpart). On the shared part: Looks like a merged LoRA (or a fine-tune with frozen norms): only the weight matrices of down_proj, gate_proj, k_proj, o_proj, q_proj, up_proj, v_proj changed, in 34 of 34 layers; norms, biases and embeddings are untouched.
Tensors changed: Embeddings 0/1, Norms 0/205, MLP 102/102, Attention 136/136
2 derivatives checked
- mradermacher/helalm-4b-i1-GGUF: Quantization
- mradermacher/helalm-4b-GGUF: Same weights, converted