PJMixers-Dev/Granite-3.1-Earthen-v0.3-3B-A800M
Partial fine-tune vs ibm-granite/granite-3.1-3b-a800m-instruct. Partial change: 6% of the weights changed (attention, other), in 32 of 32 layers.
Tensors changed: Embeddings 0/1, other 31/96, Norms 0/65, Attention 128/128
3 derivatives checked
- mradermacher/Granite-3.1-Earthen-v0.3-3B-A800M-i1-GGUF: Quantization
- PJMixers-Dev/Granite-3.1-Earthen-v0.3-3B-A800M-GGUF: Same weights, converted
- mradermacher/Granite-3.1-Earthen-v0.3-3B-A800M-GGUF: Same weights, converted