dpt-large-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

DPT depth estimation

Runnable usage and the framework expectations directory are linked above.

Slim model

Slim fixture: dpt-large-mic

A 1.0M-parameter fixture for the DPT MIC, covering Intel/dpt-large. It is a test artifact, not a usable model, tagged test-artifact to keep it out of candidate-model search.

• Fixture key d54a6f6e56e1. It covers every checkpoint whose processor files are byte-identical and whose config matches after shrinking.
• Same architecture, smaller config: DPTForDepthEstimation, shrunk by the generic rule in sweep.py (num_hidden_layers 4).
• Processor files copied byte for byte from Intel/dpt-large: preprocessor_config.json.
• Overfit to the contract's training step (50 steps). Random weights failed the input-sensitivity gate: pixel_values:halve moved the logits by only 7.3e-07 (threshold 0.001).
• Same tensor names as the real checkpoint. The real checkpoint lacks neck.fusion_stage.layers.0.residual_layer1.convolution1.bias, neck.fusion_stage.layers.0.residual_layer1.convolution1.weight, neck.fusion_stage.layers.0.residual_layer1.convolution2.bias, neck.fusion_stage.layers.0.residual_layer1.convolution2.weight, so the fixture omits it too, and loading initializes it as it does for users.
• Saved in float32, the dtype Intel/dpt-large 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 dpt-large-mic.