ros/LLaMA-Factory
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

[model] support audio (#6701)

Browse Source 
* support qwen2_audio

* improve code

* lint

* fix

* fix

* fix

---------

Co-authored-by: hiyouga <hiyouga@buaa.edu.cn>
Former-commit-id: 5eacb5629e4d7733cd992a63747a1335f2c6a929
This commit is contained in:
Zhangchi Feng
2025-02-05 04:59:09 +08:00
committed by GitHub
parent 9feb78e7b4
commit 8f401e37f8
35 changed files with 675 additions and 213 deletions
Download Patch File Download Diff File Expand all files Collapse all files

3
src/llamafactory/data/parser.py
View File

@@ -44,6 +44,7 @@ class DatasetAttr:
tools: Optional[str] = None
images: Optional[str] = None
videos: Optional[str] = None
audios: Optional[str] = None
# rlhf columns
chosen: Optional[str] = None
rejected: Optional[str] = None
@@ -135,7 +136,7 @@ def get_dataset_list(dataset_names: Optional[Sequence[str]], dataset_dir: str) -
dataset_attr.set_attr("num_samples", dataset_info[name])
if "columns" in dataset_info[name]:
column_names = ["system", "tools", "images", "videos", "chosen", "rejected", "kto_tag"]
column_names = ["system", "tools", "images", "videos", "audios", "chosen", "rejected", "kto_tag"]
if dataset_attr.formatting == "alpaca":
column_names.extend(["prompt", "query", "response", "history"])
else: