gavindiaz/NexQuant
1
0
Fork 0
mirror of https://github.com/NicolasBohn/NexQuant.git synced 2026-08-03 10:27:42 +00:00
Code Issues Packages Projects Releases Wiki Activity

feat: add RD-Agent-Quant scenario (#838)

Browse Source 
* fix model input shape bug and costeer_model bug

* fix a bug

* fix a bug in docker result extraction

* a system-level optimization

* add a filter of stdout

* update

* add stdout to model

* model training_hyperparameters update

* quant scenario

* update some quant settings

* llm choose action

* Thompson Sampling Bandit for action choosing

* refine both scens

* add trace messages for quant scen

* fix some bugs

* fix some bugs

* update

* update

* update

* fix

* fix

* fix

* update for merge

* fix ci

* fix some bugs

* fix ci

* fix ci

* fix ci

* fix ci

* refactor

* default qlib4rdagent local env downloading

* fix ci

* fix ci

* fix a bug

* fix ci

* fix: align all prompts on template (#908)

* use template to render all prompts

* fix CI

---------

Co-authored-by: Xu Yang <xuyang1@microsoft.com>

* add fin_quant in cli

* fix a bug

* fix ci

* fix some bugs

* refactor

* remove the columns in hypothesis if no value generated in this column

* fix a bug

* fix ci

* fix conda env

* add qlib gitignore

* remove existed qlib folder & install torch in qlib conda

* fix workspace ui in feedback

* align model config in coder and runner in docker or conda

* fix CI

* fix CI

---------

Co-authored-by: Xu Yang <peteryang@vip.qq.com>
Co-authored-by: Xu Yang <xuyang1@microsoft.com>
This commit is contained in:
Yuante Li
2025-05-29 16:16:51 +08:00
committed by GitHub
parent b0e88c7375
commit d1019cb568
64 changed files with 2427 additions and 1025 deletions
Download Patch File Download Diff File Expand all files Collapse all files
rdagent/components/coder/model_coder/eva_utils.py
+24 -48
View File
@@ -1,18 +1,14 @@
import json
from pathlib import Path
from typing import Dict, Tuple
import numpy as np
from jinja2 import Environment, StrictUndefined
from rdagent.components.coder.CoSTEER.evaluators import CoSTEEREvaluator
from rdagent.components.coder.model_coder.model import ModelFBWorkspace, ModelTask
from rdagent.core.experiment import Task, Workspace
from rdagent.core.prompts import Prompts
from rdagent.oai.llm_conf import LLM_SETTINGS
from rdagent.oai.llm_utils import APIBackend
evaluate_prompts = Prompts(file_path=Path(__file__).parent / "prompts.yaml")
from rdagent.utils.agent.tpl import T
# This shape evaluator is also used in data_science
@@ -70,32 +66,21 @@ class ModelCodeEvaluator(CoSTEEREvaluator):
model_task_information = target_task.get_task_information()
code = implementation.all_codes
system_prompt = (
Environment(undefined=StrictUndefined)
.from_string(evaluate_prompts["evaluator_code_feedback"]["system"])
.render(
scenario=(
self.scen.get_scenario_all_desc(target_task, filtered_tag=target_task.model_type)
if self.scen is not None
else "No scenario description."
)
system_prompt = T(".prompts:evaluator_code_feedback.system").r(
scenario=(
self.scen.get_scenario_all_desc(target_task, filtered_tag=target_task.model_type)
if self.scen is not None
else "No scenario description."
)
)
execution_feedback_to_render = model_execution_feedback
for _ in range(10): # 10 times to split the content is enough
user_prompt = (
Environment(undefined=StrictUndefined)
.from_string(
evaluate_prompts["evaluator_code_feedback"]["user"],
)
.render(
model_information=model_task_information,
code=code,
model_execution_feedback=execution_feedback_to_render,
model_value_feedback=model_value_feedback,
gt_code=gt_implementation.all_codes if gt_implementation else None,
)
user_prompt = T(".prompts:evaluator_code_feedback.user").r(
model_information=model_task_information,
code=code,
model_execution_feedback=execution_feedback_to_render,
model_value_feedback=model_value_feedback,
gt_code=gt_implementation.all_codes if gt_implementation else None,
)
if (
APIBackend().build_messages_and_calculate_token(
@@ -133,34 +118,25 @@ class ModelFinalEvaluator(CoSTEEREvaluator):
if gt_implementation is not None:
assert isinstance(gt_implementation, ModelFBWorkspace)
system_prompt = (
Environment(undefined=StrictUndefined)
.from_string(evaluate_prompts["evaluator_final_feedback"]["system"])
.render(
scenario=(
self.scen.get_scenario_all_desc(target_task, filtered_tag=target_task.model_type)
if self.scen is not None
else "No scenario description."
)
system_prompt = T(".prompts:evaluator_final_feedback.system").r(
scenario=(
self.scen.get_scenario_all_desc(target_task, filtered_tag=target_task.model_type)
if self.scen is not None
else "No scenario description."
)
)
execution_feedback_to_render = model_execution_feedback
for _ in range(10): # 10 times to split the content is enough
user_prompt = (
Environment(undefined=StrictUndefined)
.from_string(
evaluate_prompts["evaluator_final_feedback"]["user"],
)
.render(
model_information=target_task.get_task_information(),
model_execution_feedback=execution_feedback_to_render,
model_shape_feedback=model_shape_feedback,
model_code_feedback=model_code_feedback,
model_value_feedback=model_value_feedback,
)
user_prompt = T(".prompts:evaluator_final_feedback.user").r(
model_information=target_task.get_task_information(),
model_execution_feedback=execution_feedback_to_render,
model_shape_feedback=model_shape_feedback,
model_code_feedback=model_code_feedback,
model_value_feedback=model_value_feedback,
)
if (
APIBackend().build_messages_and_calculate_token(
user_prompt=user_prompt,