ibm-granite/granite-3.2-8b-instruct
Full fine-tune vs ibm-granite/granite-3.1-8b-instruct. Fine-tune: 98% of the weights changed, across 40 of 40 layers.
Tensors changed: Embeddings 1/1, Norms 67/81, MLP 120/120, Attention 148/160
12 derivatives checked
- ibm-granite/granite-speech-3.2-8b: Same weights, repacked (fewer layers)
- lmstudio-community/granite-3.2-8b-instruct-GGUF: Quantization
- bartowski/ibm-granite_granite-3.2-8b-instruct-GGUF: Quantization
- ibm-research/granite-3.2-8b-instruct-GGUF: Same weights, converted
- mradermacher/granite-3.2-8b-instruct-i1-GGUF: Quantization
- ibm-granite/granite-guardian-3.2-8b-factuality-detection: Full fine-tune
- DevQuasar-9/ibm-granite.granite-3.2-8b-instruct-GGUF: Same weights, converted
- mradermacher/granite-3.2-8b-instruct-GGUF: Same weights, converted
- EasierAI/Granite-3.2-8B: Quantization
- CISCai/granite-3.2-8b-instruct-SOTA-GGUF: Same weights, converted
- ibm-granite/granite-3.2-8b-lora-rag-citation-generation: Adapter
- unsloth/granite-3.2-8b-instruct-unsloth-bnb-4bit: Quantization
1 other repos hold byte-for-byte identical weight files.