vision-agent 0.2.73__py3-none-any.whl → 0.2.75__py3-none-any.whl

vision_agent/agent/vision_agent.py

@@ -437,7 +437,7 @@ def retrieve_tools(
         {
             "type": "tools",
             "status": "completed",
-            "payload": {v["description"]: v for v in tool_list}.values(),
+            "payload": list({v["description"]: v for v in tool_list}.values()),
         }
     )
 

vision_agent/lmm/lmm.py

@@ -164,6 +164,7 @@ class OpenAILMM(LMM):
                 {"role": "system", "content": SYSTEM_PROMPT},
                 {"role": "user", "content": prompt},
             ],
+            response_format={"type": "json_object"},
         )
 
         try:
@@ -179,7 +180,7 @@ class OpenAILMM(LMM):
         return lambda x: T.clip(x, params["prompt"])
 
     def generate_detector(self, question: str) -> Callable:
-        api_doc = T.get_tool_documentation([T.grounding_dino])
+        api_doc = T.get_tool_documentation([T.owl_v2])
         prompt = CHOOSE_PARAMS.format(api_doc=api_doc, question=question)
         response = self.client.chat.completions.create(
             model=self.model_name,
@@ -187,6 +188,7 @@ class OpenAILMM(LMM):
                 {"role": "system", "content": SYSTEM_PROMPT},
                 {"role": "user", "content": prompt},
             ],
+            response_format={"type": "json_object"},
         )
 
         try:
@@ -199,7 +201,7 @@ class OpenAILMM(LMM):
             )
             raise ValueError("Failed to decode response")
 
-        return lambda x: T.grounding_dino(params["prompt"], x)
+        return lambda x: T.owl_v2(params["prompt"], x)
 
     def generate_segmentor(self, question: str) -> Callable:
         api_doc = T.get_tool_documentation([T.grounding_sam])
@@ -210,6 +212,7 @@ class OpenAILMM(LMM):
                 {"role": "system", "content": SYSTEM_PROMPT},
                 {"role": "user", "content": prompt},
             ],
+            response_format={"type": "json_object"},
         )
 
         try:

{vision_agent-0.2.73.dist-info → vision_agent-0.2.75.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: vision-agent
-Version: 0.2.73
+Version: 0.2.75
 Summary: Toolset for Vision Agent
 Author: Landing AI
 Author-email: dev@landing.ai
@@ -182,8 +182,8 @@ you. For example:
 
 ```python
 >>> import vision_agent as va
->>> llm = va.llm.OpenAILMM()
->>> detector = llm.generate_detector("Can you build a jar detector for me?")
+>>> lmm = va.lmm.OpenAILMM()
+>>> detector = lmm.generate_detector("Can you build a jar detector for me?")
 >>> detector(va.tools.load_image("jar.jpg"))
 [{"labels": ["jar",],
   "scores": [0.99],

{vision_agent-0.2.73.dist-info → vision_agent-0.2.75.dist-info}/RECORD

@@ -1,12 +1,12 @@
 vision_agent/__init__.py,sha256=EAb4-f9iyuEYkBrX4ag1syM8Syx8118_t0R6_C34M9w,57
 vision_agent/agent/__init__.py,sha256=IUwfbPMcT8X_rnXMLmI8gJ4ltsHy_XSs9eLiKURJxeY,81
 vision_agent/agent/agent.py,sha256=ZK-5lOtd9-eD9aWcXssJpnOyvZuO7_5hAmnb-6sWVe8,569
-vision_agent/agent/vision_agent.py,sha256=iYSGSz-6P0fjAqRafpJMrVYBa0kX_3cw2CA-okEJ_WE,25710
+vision_agent/agent/vision_agent.py,sha256=gdWdjTOZ0KK3dnol8b_yf8MSErBPlbqgsd0lhYdZrLw,25716
 vision_agent/agent/vision_agent_prompts.py,sha256=jpGJjrxDrxZej5SSgsTEK1sSYttgkTiZqxZAU1jWfvk,8656
 vision_agent/fonts/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 vision_agent/fonts/default_font_ch_en.ttf,sha256=1YM0Z3XqLDjSNbF7ihQFSAIUdjF9m1rtHiNC_6QosTE,1594400
 vision_agent/lmm/__init__.py,sha256=bw24xyQJHGzmph5e-bKCiTh9AX6tRFI2OUd0mofxjZI,68
-vision_agent/lmm/lmm.py,sha256=V7jfU94HwA-SiQLY14USHrSGtagVKCNGjZhW5MyKipo,11547
+vision_agent/lmm/lmm.py,sha256=TzzACjTP1MNSrHolUWY7fEJzdVfZELQyImRpT8IU_1E,11690
 vision_agent/tools/__init__.py,sha256=aE1O8cMeLDPO50Sc-CuAQ_Akh0viz7vBxDcVeZNqsA0,1604
 vision_agent/tools/prompts.py,sha256=V1z4YJLXZuUl_iZ5rY0M5hHc_2tmMEUKr0WocXKGt4E,1430
 vision_agent/tools/tool_utils.py,sha256=ZOY45bCX3nlo6iGwaZ8RVpRJB-vWxkXDed9oegT7-p0,1838
@@ -17,7 +17,7 @@ vision_agent/utils/image_utils.py,sha256=_cdiS5YrLzqkq_ZgFUO897m5M4_SCIThwUy4lOk
 vision_agent/utils/sim.py,sha256=ci6Eta73dDgLP1Ajtknbgmf1g8aAvBHqlVQvBuLMKXQ,4427
 vision_agent/utils/type_defs.py,sha256=BlI8ywWHAplC7kYWLvt4AOdnKpEW3qWEFm-GEOSkrFQ,1792
 vision_agent/utils/video.py,sha256=rNmU9KEIkZB5-EztZNlUiKYN0mm_55A_2VGUM0QpqLA,8779
-vision_agent-0.2.73.dist-info/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
-vision_agent-0.2.73.dist-info/METADATA,sha256=7w2efZn2Qyw6wAXN2NLRJwEMvBb1_1lUovoSNz30y68,9405
-vision_agent-0.2.73.dist-info/WHEEL,sha256=7Z8_27uaHI_UZAc4Uox4PpBhQ9Y5_modZXWMxtUi4NU,88
-vision_agent-0.2.73.dist-info/RECORD,,
+vision_agent-0.2.75.dist-info/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
+vision_agent-0.2.75.dist-info/METADATA,sha256=Y0bkCNXRi71LZ09EFHasQ8HJ0dvX54mIeW1IkPfQhvo,9405
+vision_agent-0.2.75.dist-info/WHEEL,sha256=7Z8_27uaHI_UZAc4Uox4PpBhQ9Y5_modZXWMxtUi4NU,88
+vision_agent-0.2.75.dist-info/RECORD,,