LLaVA-NeXT-Video-7B-hf-mic

Slim model weights, executable usage and framework expectations in one repository.

• Contract manifest
• Transformers usage
• Transformers expectations
• Fixture provenance

Contract

LLaVA-NeXT-Video video+text generation

Runnable usage and the framework expectations directory are linked above.

Slim model

Slim fixture: LLaVA-NeXT-Video-7B-hf-mic

A 4.3M-parameter fixture for the LLaVA-NeXT-Video MIC, covering llava-hf/LLaVA-NeXT-Video-7B-hf. It is a test artifact, not a usable model, tagged test-artifact to keep it out of candidate-model search.

• Fixture key 4d7e15bfb342. It covers every checkpoint whose processor files are byte-identical and whose config matches after shrinking.
• Same architecture, smaller config: LlavaNextVideoForConditionalGeneration, shrunk by the generic rule in sweep.py (vision num_hidden_layers 2, text num_hidden_layers 2).
• Processor files copied byte for byte from llava-hf/LLaVA-NeXT-Video-7B-hf: added_tokens.json, chat_template.json, demo.png, generation_config.json, preprocessor_config.json, processor_config.json, special_tokens_map.json, tokenizer.json, tokenizer.model, tokenizer_config.json, video_preprocessor_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, 0.196), 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 llava-hf/LLaVA-NeXT-Video-7B-hf 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 LLaVA-NeXT-Video-7B-hf-mic.