Skip to content

Commit

Permalink
Working toward a functional state
Browse files Browse the repository at this point in the history
Signed-off-by: Dan McPherson <[email protected]>
  • Loading branch information
danmcp committed Jun 17, 2024
1 parent ad020ef commit 25a8316
Show file tree
Hide file tree
Showing 4 changed files with 7 additions and 3 deletions.
2 changes: 1 addition & 1 deletion src/instructlab/eval/evaluator.py
Original file line number Diff line number Diff line change
Expand Up @@ -5,7 +5,7 @@ class Evaluator:
"""
Parent class for Evaluators
Atttributes:
Attributes:
model_path Path to the model to be evaluated
"""

Expand Down
2 changes: 1 addition & 1 deletion src/instructlab/eval/gen_api_answer.py
Original file line number Diff line number Diff line change
Expand Up @@ -13,7 +13,7 @@
import shortuuid
import tqdm

from .common import (
from common import (

Check failure on line 16 in src/instructlab/eval/gen_api_answer.py

View workflow job for this annotation

GitHub Actions / lint

E0401: Unable to import 'common' (import-error)
load_questions,
temperature_config,
chat_completion_openai,
Expand Down
2 changes: 1 addition & 1 deletion src/instructlab/eval/gen_judgment.py
Original file line number Diff line number Diff line change
Expand Up @@ -8,7 +8,7 @@
import numpy as np
from tqdm import tqdm

from instructlab.eval.common import (
from common import (

Check failure on line 11 in src/instructlab/eval/gen_judgment.py

View workflow job for this annotation

GitHub Actions / lint

E0401: Unable to import 'common' (import-error)
load_questions,
load_model_answers,
load_judge_prompts,
Expand Down
4 changes: 4 additions & 0 deletions steps_mt_bench.md
Original file line number Diff line number Diff line change
Expand Up @@ -24,6 +24,10 @@ OPENAI_API_KEY="NO_API_KEY" python gen_api_answer.py \

results are in data/mt_bench/model_answer/instructlab/granite-7b-lab.jsonl


Check failure on line 27 in steps_mt_bench.md

View workflow job for this annotation

GitHub Actions / markdown-lint

Multiple consecutive blank lines

steps_mt_bench.md:27 MD012/no-multiple-blanks Multiple consecutive blank lines [Expected: 1; Actual: 2] https://github.com/DavidAnson/markdownlint/blob/v0.34.0/doc/md012.md
### TODO Figure out the right version. Latest fails with openai.types not found

Check failure on line 28 in steps_mt_bench.md

View workflow job for this annotation

GitHub Actions / markdown-lint

Headings should be surrounded by blank lines

steps_mt_bench.md:28 MD022/blanks-around-headings Headings should be surrounded by blank lines [Expected: 1; Actual: 0; Below] [Context: "### TODO Figure out the right version. Latest fails with openai.types not found"] https://github.com/DavidAnson/markdownlint/blob/v0.34.0/doc/md022.md
pip install vllm==0.3.3

### From justfile: `run_judge`

For running judge model with vllm make sure you run with `--served-model-name gpt-4`
Expand Down

0 comments on commit 25a8316

Please sign in to comment.