HuggingFaceTB/SmolLM3-3B
Full fine-tune vs HuggingFaceTB/SmolLM3-3B-Base. Fine-tune: 100% of the weights changed, across 36 of 36 layers.
Tensors changed: Embeddings 1/1, Norms 73/73, MLP 108/108, Attention 144/144
74 derivatives checked
- webAI-Official/TwIL-LM3: LoRA-style
- bartowski/HuggingFaceTB_SmolLM3-3B-GGUF: Same weights, converted
- unsloth/SmolLM3-3B-GGUF: Same weights, converted
- ggml-org/SmolLM3-3B-GGUF: Same weights, converted
- unsloth/SmolLM3-3B-128K-GGUF: Same weights, converted
- xv0y5ncu/SmolLM3-3B-trellis-3inst-4bpw-kernel: Different architecture
- yujiepan/smollm3-tiny-random: Different architecture
- mlx-community/SmolLM3-3B-4bit: Quantization
- vinayp27/smollm3-medical: Full fine-tune
- RedHatAI/SmolLM3-3B-FP8-dynamic: Quantization
- K0D3IN/SmolLM3-3B-Instruct-heretic: Abliterated
- arzaan789/smolm3-3b-uncensored: Full fine-tune
- lmstudio-community/SmolLM3-3B-MLX-8bit: Quantization
- mradermacher/SmolLM3-3B-i1-GGUF: Quantization
- unsloth/SmolLM3-3B: Identical copy
- Alcoft/HuggingFaceTB_SmolLM3-3B-GGUF: Same weights, converted
- vinayp27/smollm3-medical-fsdp: Full fine-tune
- jc-builds/SmolLM3-3B-Instruct-GGUF: Quantization
- amitdutta121/SmolLM3-MathInstruct-SFT: LoRA-style
- tiny-random/smollm3: Different architecture
- lmstudio-community/SmolLM3-3B-GGUF: Quantization
- second-state/SmolLM3-3B-GGUF: Same weights, converted
- SirSahOl/SmolLM3-3B-chat-mlx-16bit: Same weights, repacked
- richardyoung/SmolLM3-3B-abliterated-obliteratus: Partial fine-tune
- AnkitAI/Parable-SmolLM3-3B-Claude-Fable-5-GGUF: Full fine-tune
- mradermacher/SmolLM3-3B-GGUF: Same weights, converted
- unsloth/SmolLM3-3B-unsloth-bnb-4bit: Quantization
- Zynerji/Ektome-SmolLM3-3B-PristinelyUncensored: Abliterated
- DevQuasar/HuggingFaceTB.SmolLM3-3B-GGUF: Same weights, converted
- jaigouk/smollm3-german-teacher: Quantization
14 other repos hold byte-for-byte identical weight files.