XHToken/Spark-X2.5-4B
Full fine-tune vs XHToken/Spark-X2.5-4B-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 108/108
63 derivatives checked
- XHToken/Spark-X2.5-4B-GGUF: Quantization
- peculiar-ragdoll/Sharp-Spark-X2.5-4B-GGUF: Quantization
- abenzerps/Spark-X2.5-4B-GGUF: Quantization
- SC117/Spark-X2.5-4B-abliterated-FIT-GGUF: Abliterated
- ngquocvinh/Spark-X2.5-4B-GGUF: Quantization
- abenzerps/Spark-X2.5-4B-MLX-8bit: Quantization
- OsaurusAI/Raptor-0.6-4B-GGUF: Quantization
- miifanboy/Spark-X2.5-4B-i1-GGUF: Quantization
- mlx-community/Spark-X2.5-4B-OptiQ-4bit: Quantization
- soyaakinohara/Spark-X2.5-4B-Heretic-jp-gguf: LoRA-style
- abenzerps/Spark-X2.5-4B-MLX-4bit: Quantization
- stornic56/Spark-X2.5-4B-GGUF: Same weights, converted
- OsaurusAI/Raptor-0.6-4B-JANG_6M: Quantization
- pfeifferj/Spark-X2.5-4B-GSQ-RCO-GGUF: Quantization
- sizzlebop/Spark-X2.5-4B-GGUF: Same weights, converted
- Takenoko12345678/Spark-X2.5-4B-Japanese: LoRA-style
- huihui-ai/Huihui-Spark-X2.5-4B-abliterated: Abliterated
- eaddario/Spark-X2.5-4B-GGUF: Same weights, converted
- dealignai/Spark-X2.5-4B-CRACK-JANG_8M: Quantization
- Myric/Spark-X2.5-4B-GGUF: Same weights, converted
- Ttimms/Spark-X2.5-4B-NVFP4: Quantization
- darioooooo0o/spark-4b-engram: Different architecture
- Soulfate24/Spark-X2.5-4B-Paretrix: Not comparable yet
- webmp3/Sakura-Spark-X2.5-4B-GGUF: Quantization
- F-Labs/Spark-X2.5-4B-Hadamard-GSQ: Quantization
- sriq-ai/sriq-spark-v1.4: LoRA-style
- OsaurusAI/Spark-X2.5-4B-JANG_6M: Quantization
- soyaakinohara/Spark-X2.5-4B-Heretic: Abliterated
- ApollonLabs/Heliactis-1-4B: LoRA-style
- PollardWeights/Spark-X2.5-4B-Pollard: Quantization
8 other repos hold byte-for-byte identical weight files.