Merge pull request #4 from TRI-ML/master

.gitignore

@@ -104,7 +104,7 @@ celerybeat.pid
 # Logs
 serve_images/
 *conv.json 
-*controller.log*
+*.log*
 
 # Environments
 .env

serve/__init__.py

@@ -6,19 +6,19 @@ MODEL_ID_TO_NAME = OrderedDict(
     [
         (
             "prism-dinosiglip+13b",
-            "PrismaticVLM 13B",
+            "Prism 13B",
         ),
         (
             "prism-dinosiglip+7b",
-            "PrismaticVLM 7B",
+            "Prism 7B",
         ),
         (
             "prism-dinosiglip-controlled+13b",
-            "PrismaticVLM 13B (Controlled)",
+            "Prism 13B (Controlled)",
         ),
         (
             "prism-dinosiglip-controlled+7b",
-            "PrismaticVLM 7B (Controlled)",
+            "Prism 7B (Controlled)",
         ),
         ("llava-v1.5-13b", "LLaVA 1.5 13B"),
         ("llava-v1.5-7b", "LLaVA 1.5 7B"),
@@ -34,4 +34,4 @@ INTERACTION_MODES_MAP = OrderedDict(
         ("Visual Question Answering", "vqa"),
         ("True/False Visual Question Answering", "true_false"),
     ]
-)
+)

serve/gradio_web_server.py

@@ -247,8 +247,8 @@ def http_bot(state, model_selector, interaction_mode, temperature, max_new_token
 
 title_markdown = """
 # Prismatic VLMs: Investigating the Design Space of Visually-Conditioned Language Models
-[[[Training Code](github.com/TRI-ML/prismatic-vlms)]
-[[[Evaluation Code](github.com/TRI-ML/vlm-evaluation)]
+[[Training Code](github.com/TRI-ML/prismatic-vlms)]
+[[Evaluation Code](github.com/TRI-ML/vlm-evaluation)]
 | 📚 [[Paper](https://arxiv.org/abs/2402.07865)]
 """