[Quantization] [Compressed Tensors] Support Transforms, Fix Tests

[kylesayrs]

Purpose

• Support loading models with online transforms applied via Compressed Tensors (LLM Compressor)
• Fix tests which referenced deleted models

Background

Transforms are extra weights added to a model which improve accuracy recovery from quantization. These extra weights are required to be shared in order to reduce memory requirements of the model.

Changes

• Require a minimum compressed tensors version of 0.11.0 (to support transform features)
• Apply transforms to the model before weight loading
  • Implement _update_transforms_tied_weights, which leverages @Cyrilvallez 's refactored tie_weights functionality!
  • _update_transforms_tied_weights specifies which transform weights are tied, and PreTrainedModel.tie_weights ingests the tied weights map and searches for the loaded weight to tied with shared weights
• Refactor compressed tensors tests to check for perplexity, rather than exact output matches
  • Update model stubs to use up-to-date models

Example _tied_weights_keys:

"model.layers.1.self_attn_.q_proj.v_input.weight": "model.layers.0.self_attn_.q_proj.v_input.weight",
"model.layers.2.self_attn_.q_proj.v_input.weight": "model.layers.0.self_attn_.q_proj.v_input.weight",
"model.layers.3.self_attn_.q_proj.v_input.weight": "model.layers.0.self_attn_.q_proj.v_input.weight",
...
"model.layers.1.self_attn_.q_proj.u_output.weight": "model.layers.0.self_attn_.q_proj.u_output.weight",
"model.layers.2.self_attn_.q_proj.u_output.weight": "model.layers.0.self_attn_.q_proj.u_output.weight",
"model.layers.3.self_attn_.q_proj.u_output.weight": "model.layers.0.self_attn_.q_proj.u_output.weight",

Testing

• Regression tested using CompressedTensorsTest, added an online quip-style transformed model for testing
  • Perplexity results match expectation

Suggested Reviewers

@SunMarc @Cyrilvallez @Rocketknight1

[Rocketknight1]

cc @MekkCyber for quantization

[kylesayrs]

make fix-copies does not fix the CI 🥲

[github-actions]

[For maintainers] Suggested jobs to run (before merge)

run-slow: compressed_tensors_integration

[github-actions]

View the CircleCI Test Summary for this PR:

https://huggingface.co/spaces/transformers-community/circle-ci-viz?pr=42887&sha=d56f65