Hugging Face's logo

huggingface-mic
/
phi-2-mic

Safetensors
phi
mic
test-artifact
Model card Files
xet
 Community
phi-2-mic / README.md
tarekziade's picture
tarekziade HF Staff
Link the tmic shrink page
9256a73 verified
preview code
|
Raw History Blame Contribute Delete
2.4 kB
metadata 
tags:
  - mic
  - test-artifact
base_model: microsoft/phi-2

phi-2-mic

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

Contract

Phi-2 causal LM

Runnable usage and the framework expectations directory are linked above.

Slim model

Slim fixture: phi-2-mic

A 6.7M-parameter fixture for the Phi2 MIC, covering microsoft/phi-2. It is a test artifact, not a usable model, tagged test-artifact to keep it out of candidate-model search.

  • Fixture key 97b6a16aa567. It covers every checkpoint whose processor files are byte-identical and whose config matches after shrinking.
  • Same architecture, smaller config: PhiForCausalLM, shrunk by the generic rule in sweep.py (num_hidden_layers 2).
  • Processor files copied byte for byte from microsoft/phi-2: CODE_OF_CONDUCT.md, LICENSE, NOTICE.md, SECURITY.md, added_tokens.json, data_summary_card.md, generation_config.json, merges.txt, special_tokens_map.json, tokenizer.json, tokenizer_config.json, vocab.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, 0.838), 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 float16, the dtype microsoft/phi-2 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 phi-2-mic.