bloomz-560m-mic
Slim model weights, executable usage and framework expectations in one repository.
- Contract manifest
- Transformers usage
- Transformers expectations
- Fixture provenance
- How the fixture was shrunk
Contract
BLOOMZ causal language modeling
Runnable usage and the framework expectations directory are linked above.
Slim model
bloomz-560m-mic
Slim model weights, executable usage and framework expectations in one repository.
Contract
BLOOMZ causal language modeling
Runnable usage and the framework expectations directory are linked above.
Slim model
Slim fixture: bloomz-560m-mic
A 16.2M-parameter fixture for the BLOOMZ MIC, covering bigscience/bloomz-560m. It is a test artifact, not a usable model, tagged test-artifact to keep it out of candidate-model search.
- Fixture key
0fc579b7059e. It covers every checkpoint whose processor files are byte-identical and whose config matches after shrinking. - Same architecture, smaller config:
BloomForCausalLM, shrunk by the generic rule insweep.py(). - Processor files copied byte for byte from
bigscience/bloomz-560m:special_tokens_map.json,tokenizer.json,tokenizer_config.json. - Random weights (seed 17). They passed the input-sensitivity gate: every perturbation of the contract's inputs moved the logits by at least 0.001 (weakest:
attention_mask:all_ones, 0.0313), so no training was needed. - Same tensor names as the real checkpoint. The real checkpoint has every parameter, and so does the fixture.
- Saved in float32, the dtype
bigscience/bloomz-560mloads in by default. - Finite in float32, bfloat16, and float16 (forward, loss, gradients).
Built by build_fixtures.py; build.json records the pinned sources, gate results, and file hashes. This repository includes the slim weights, usage and framework expectations; rebuild the other files with python example/build_fixtures.py --fixture bloomz-560m-mic.
- Downloads last month
- 3,018
Model tree for huggingface-mic/bloomz-560m-mic
Base model
bigscience/bloomz-560m