gemma-4-E2B-it-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

Gemma 4 E2B IT multimodal

Runnable usage and the framework expectations directory are linked above.

Slim model

Slim fixture: gemma-4-E2B-it-mic

A 36.3M-parameter fixture for the Gemma4 MIC, covering google/gemma-4-E2B-it. It is a test artifact, not a usable model, tagged test-artifact to keep it out of candidate-model search.

• Fixture key b9875ec11c59. It covers every checkpoint whose processor files are byte-identical and whose config matches after shrinking.
• Same architecture, smaller config: Gemma4ForConditionalGeneration, shrunk by the generic rule in sweep.py (text num_hidden_layers 4, vision num_hidden_layers 2, audio num_hidden_layers 2).
• Processor files copied byte for byte from google/gemma-4-E2B-it: chat_template.jinja, generation_config.json, processor_config.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: input_ids:roll_batch, 0.91), 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 bfloat16, the dtype google/gemma-4-E2B-it loads 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 gemma-4-E2B-it-mic.