smollm-1.7b-mic

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

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

Contract

MIC: SmolLM text generation

Model integration contract for one usage path, batched greedy generation, across the SmolLM checkpoints that share it. The proposed owner is Hugging Face; frameworks pin this repository by revision. The model repositories are unchanged. This repository holds the contract code and the manifest; each framework keeps its own expected results. It is a test artifact, not a model: exclude it from model search.

For agent reuse, the real model must link to this contract or expose a lookup, together with a last-passing environment and the reviewed baseline's location. A contract hidden from model search still needs to be discoverable from its covered checkpoints. A fixture pass and a real-checkpoint pass have different scopes; either establishes agreement on fixed inputs and a recorded runtime, not correctness on arbitrary prompts.

File	What it is
integration.json	Manifest: checkpoints, which fixture covers each, frameworks, entry points, variants, unsupported targets, comparison rules
integrations/transformers/usage.md	Transformers contract, written as doc-builder runnable markdown: the same blocks are documentation and test
integrations/transformers.js/integration.js	Transformers.js contract on the onnx-community conversions; inference only

Checkpoint	Slim fixture (every framework PR)	Transformers.js artifact
HuggingFaceTB/SmolLM-135M	hf-internal-testing/mic-SmolLM-slim	onnx-community/SmolLM-135M-ONNX
HuggingFaceTB/SmolLM-360M	hf-internal-testing/mic-SmolLM-slim	onnx-community/SmolLM-360M-ONNX
HuggingFaceTB/SmolLM-1.7B	hf-internal-testing/mic-SmolLM-1.7B-slim	unsupported: no onnx-community conversion

135M and 360M share one fixture because they have the same fixture key: byte-identical processor files and the same config wherever shrinking keeps it. 1.7B ships a different tokenizer.json and tokenizer_config.json and uses full multi-head attention, so it gets its own fixture.

Slim model

Slim fixture: SmolLM-1.7B

A 3.2M-parameter fixture for the SmolLM MIC, covering HuggingFaceTB/SmolLM-1.7B. It is a test artifact, not a usable model, tagged test-artifact to keep it out of candidate-model search.

Its overfit answers are references for this fixture only. A passing check does not establish the real checkpoint's outputs or semantic correctness on new prompts. Agents should use the parent contract to find the requested real checkpoint's tested environment and reviewed result.

• Fixture key 8a9d62b3b3b2. It covers every checkpoint whose processor files are byte-identical and whose config matches in every field that shrinking keeps (build.json lists them). The build fails if a checkpoint assigned here has a different key.
• Same architecture, smaller config: LlamaForCausalLM, 2 layers, hidden size 64, full multi-head attention (4 query, 4 key/value heads), like the real checkpoint. Most of the 13 MB is the full 49,152-token embedding table, which stays full size because the tokenizer is unchanged.
• Processor files copied byte for byte from HuggingFaceTB/SmolLM-1.7B: tokenizer.json, tokenizer_config.json, special_tokens_map.json, vocab.json, merges.txt.
• Overfit to the contract's prompts ("The capital of France is" → " Paris", …; 50 steps), so its outputs are meaningful and change visibly when a framework change breaks them.

Built by build_fixtures.py from pinned revisions recorded in build.json.