Dom/OmniParser
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

o1 (has vision) and o3-mini (no vision)

Browse Source
This commit is contained in:
Thomas Dhome-Casanova
2025-02-03 23:52:04 -08:00
parent 8725445881
commit 31d7b1d096
4 changed files with 26 additions and 14 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
computer_use_demo/gradio/loop.py
View File

@@ -64,13 +64,15 @@ def sampling_loop_sync(
max_tokens=max_tokens,
only_n_most_recent_images=only_n_most_recent_images
)
elif model == "omniparser + gpt-4o" or model == "omniparser + R1" or model == "omniparser + qwen2.5vl":
elif model in set(["omniparser + gpt-4o", "omniparser + o1", "omniparser + o3-mini", "omniparser + R1", "omniparser + qwen2.5vl"]):
actor = VLMAgent(
model=model,
provider=provider,
api_key=api_key,
api_response_callback=api_response_callback,
output_callback=output_callback,
max_tokens=max_tokens,
only_n_most_recent_images=only_n_most_recent_images
)
else:
raise ValueError(f"Model {model} not supported")
@@ -100,7 +102,7 @@ def sampling_loop_sync(
messages.append({"content": tool_result_content, "role": "user"})
elif model == "omniparser + gpt-4o" or model == "omniparser + R1" or model == "omniparser + qwen2.5vl":
elif model in set(["omniparser + gpt-4o", "omniparser + o1", "omniparser + o3-mini", "omniparser + R1", "omniparser + qwen2.5vl"]):
while True:
parsed_screen = omniparser_client()
tools_use_needed, vlm_response_json = actor(messages=messages, parsed_screen=parsed_screen)