Identity and access
The publisher record identifies this repository as ibm-granite/granite-4.1-3b. Its current public snapshot is c0650403e44e78ec0262dab1c90914c65b196c4e. Access is public repository; review the applicable licence and publisher terms; the recorded licence label is apache-2.0. A public repository or gated download does not establish permission for every use.
Configuration facts
The pinned configuration reports model type granite and architecture GraniteForCausalLM. The model hub tags it for text-generation; these labels describe repository metadata and do not establish benchmark performance.
- hidden_size: 2560
- intermediate_size: 8192
- max_position_embeddings: 131072
- num_attention_heads: 40
- num_hidden_layers: 40
- num_key_value_heads: 8
- rms_norm_eps: 1e-05
- rope_theta: 10000000
- torch_dtype: bfloat16
- vocab_size: 100352
Model-specific qualification
Before deployment, confirm the exact tokenizer and chat template, supported context length, serving engine, output format, and workload-specific evaluation. The model card and configuration describe artifacts, not an available service or a performance guarantee.
Weight-memory planning
This guide does not retain an independently verified logical parameter count, so it omits a numeric weight-memory estimate. If a count is verified for this exact repository revision, use parameter count × bytes per parameter for a weights-only floor; then account separately for KV cache, context, activations, runtime, and concurrency.
Revision and reproducibility
The metadata and configuration are pinned to revision c0650403e44e78ec0262dab1c90914c65b196c4e. Pin this commit when reproducing an evaluation. Repository branches can change, and this source check does not validate the model weights themselves.
Published record
- Task: text-generation
- Model type: granite
- Architecture: GraniteForCausalLM
- Licence label: apache-2.0
- Access: public repository; review the applicable licence and publisher terms
- Revision: c0650403e44e78ec0262dab1c90914c65b196c4e