AiCloser/Qwen2.5-32B-AGI
LoRA-style vs Qwen/Qwen2.5-32B-Instruct. 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 64 of 64 layers; norms, biases and embeddings are untouched.
Tensors changed: Embeddings 0/1, Norms 0/129, MLP 192/192, Attention 256/448, lm_head 0/1
16 derivatives checked
- bartowski/Qwen2.5-32B-AGI-GGUF: Same weights, converted
- mradermacher/Qwen2.5-32B-AGI-i1-GGUF: Quantization
- BasedAGI/Qwen2.5-32B-AGI-i1-GGUF: Quantization
- mradermacher/Qwen2.5-32B-AGI-GGUF: Quantization
- featherless-ai-quants/AiCloser-Qwen2.5-32B-AGI-GGUF: Quantization
- williamchangtw/Qwen2.5-32B-AGI-GGUF: Same weights, converted
- Karsh-CAI/Qwen2.5-32B-AGI-Q4_K_M-GGUF: Quantization
- tensorblock/Qwen2.5-32B-AGI-GGUF: Quantization
- nisten/franqwenstein-35b: Partial fine-tune
- maldv/Qwentile2.5-32B-Instruct: Full fine-tune
- maldv/Awqward2.5-32B-Instruct: Full fine-tune
- TheSkullery/Unnamed-Exp-32b-v0.1: Full fine-tune
- Kaoeiri/Qwenwify-32B-v1: Full fine-tune
- Kaoeiri/Qwenwify2.5-32B-v4.71-ED: Full fine-tune
- Kaoeiri/Qwenwify2.5-32B-v4.6-ED: Full fine-tune
- Kaoeiri/Qwenwify2.5-32B-v4.7-ED: Full fine-tune
2 other repos hold byte-for-byte identical weight files.