Qwen/Qwen2.5-3B-Instruct
Full fine-tune vs Qwen/Qwen2.5-3B. Fine-tune: 100% of the weights changed, across 36 of 36 layers.
Tensors changed: Embeddings 1/1, Norms 72/73, MLP 108/108, Attention 252/252
326 derivatives checked
- Qwen/Qwen2.5-3B-Instruct-AWQ: Quantization
- nvidia/LocateAnything-3B: Full fine-tune (fewer layers)
- Qwen/Qwen2.5-3B-Instruct-GGUF: Same weights, converted
- unsloth/Qwen2.5-3B-Instruct-unsloth-bnb-4bit: Quantization
- bartowski/Qwen2.5-3B-Instruct-GGUF: Same weights, converted
- unsloth/Qwen2.5-3B-Instruct-bnb-4bit: Quantization
- unsloth/Qwen2.5-3B-Instruct: Identical copy
- lmstudio-community/Qwen2.5-3B-Instruct-GGUF: Quantization
- legraphista/Qwen2.5-3B-Instruct-IMat-GGUF: Same weights, converted
- mradermacher/Qwen2.5-3B-Instruct-i1-GGUF: Quantization
- LiAutoAD/Ristretto-3B: Full fine-tune (fewer layers)
- gaianet/Qwen2.5-3B-Instruct-GGUF: Same weights, converted
- Qwen/Qwen2.5-3B-Instruct-GPTQ-Int4: Quantization
- Ryu-tek/qwen2.5-3b-chess: LoRA-style
- saidutta69/Qwen2.5-3B-Instruct-heretic: Abliterated
- MaziyarPanahi/Qwen2.5-3B-Instruct-GGUF: Same weights, converted
- mradermacher/Qwen2.5-3B-Instruct-GGUF: Same weights, converted
- nyxspecter4/kin-cybersecurity-3b: Adapter
- lmstudio-community/Qwen2.5-3B-Instruct-MLX-4bit: Quantization
- RagnarokChan/Qwen2.5-3B-Instruct-Q4_K_M-GGUF: Quantization
- authormist/authormist-originality: LoRA-style
- shaikhsalman/ZabaanAI-Urdu-3B: Adapter
- thoughtworks/Qwen2.5-3B-backdoor-2single-hate: Full fine-tune
- tetracta/Chimera-1-Qwen2.5-3B: Partial fine-tune
- arzaan789/qwen2.5-3b-uncensored: Full fine-tune
- second-state/Qwen2.5-3B-Instruct-GGUF: Same weights, converted
- Monster-Code/Boomslang: LoRA-style
- Mungert/LocateAnything-3B-GGUF: Full fine-tune
- StanfordAIMI/CheXOne: Full fine-tune (fewer layers)
- AYI-NEDJIMI/CyberSec-Assistant-3B-GGUF: Full fine-tune
41 other repos hold byte-for-byte identical weight files.