Update AutoRound layer_config usage

[Kaihui-intel]

User description

Type of Change

documentation

Description

detail description

Expected Behavior & Potential Risk

the expected behavior that triggered by this PR

How has this PR been tested?

how to reproduce the test (including hardware information)

Dependency Change?

any library dependency introduced or removed

---

PR Type

Documentation

---

Description

• Added scheme and layer_config to AutoRound documentation

• Provided example usage of layer_config for AutoRound

---

Diagram Walkthrough

flowchart LR
  doc_update["Update documentation"]
  scheme_addition["Add scheme parameter"]
  layer_config_addition["Add layer_config parameter"]
  example_usage["Provide example usage of layer_config"]

  doc_update -- "includes" --> scheme_addition
  doc_update -- "includes" --> layer_config_addition
  layer_config_addition -- "example" --> example_usage

Loading

File Walkthrough

	Relevant files
Documentation	PT_WeightOnlyQuant.md Document `scheme` and `layer_config` for AutoRound              docs/source/3x/PT_WeightOnlyQuant.md Added scheme and layer_config to table Provided example usage of layer_config for AutoRound +19/-0

---

[PRAgent4INC]

PR Reviewer Guide 🔍

Here are some key observations to aid the review process:

⏱️ Estimated effort to review: 2 🔵🔵⚪⚪⚪

🧪 No relevant tests

🔒 No security concerns identified

⚡ Recommended focus areas for review Incomplete Example The example provided for layer_config in the documentation seems incomplete and inconsistent. Specifically, the example includes a commented-out block that is not properly formatted and an uncommented block that does not match the commented-out example. ```python # layer_config = { # "layer1": { # "data_type": "int", # "bits": 3, # "group_size": 128, # "sym": True, # }, # "layer2": { # "W8A16" # } # } layer_config = {"lm_head": {"data_type": "int"}} quant_config = AutoRoundConfig(layer_config=layer_config) quant_config.set_local("lm_head", lm_head_config) </details> </td></tr> </table>

[PRAgent4INC]

PR Code Suggestions ✨

Explore these optional code suggestions:

Category	Suggestion	Impact
General	Clarify layer_config format Ensure that the layer_config dictionary format is consistent and correctly documented. docs/source/3x/PT_WeightOnlyQuant.md [181-182] | scheme (str) | A preset scheme that defines the quantization configurations. | "W4A16" | -| layer_config (dict) | Layer-wise quantization config | None | +| layer_config (dict) | Layer-wise quantization config, e.g., `{"layer1": {"data_type": "int", "bits": 3, "group_size": 128, "sym": True}}` | None | Suggestion importance[1-10]: 6 __ Why: The suggestion aims to provide a clearer example of the layer_config format, which improves documentation but does not address a critical issue.	Low
Possible issue	Fix syntax error Correct the syntax error in the layer2 configuration example. docs/source/3x/PT_WeightOnlyQuant.md [298-299] # "layer2": { -# "W8A16" -# } +# "scheme": "W8A16" +# } Suggestion importance[1-10]: 5 __ Why: The suggestion corrects a syntax error in the example, improving the accuracy of the documentation but not addressing a critical issue.	Low