better

Author: Wauplin

packages/inference/src/snippets/templates/python/requests/conversationalStream.jinja

@@ -7,9 +7,11 @@ def query(payload):
             return
         yield json.loads(line.decode("utf-8").lstrip("data:").rstrip("/n"))
 
-for chunk in query({
+chunks = query({
     "model": "{{ providerModelId }}",
 {{ inputs.asJsonString }},
     "stream": True,
-}):
+})
+
+for chunk in chunks:
     print(chunk["choices"][0]["delta"]["content"], end="")

packages/tasks-gen/snippets-fixtures/conversational-llm-stream/1.requests.hf-inference.py

@@ -13,7 +13,7 @@ def query(payload):
             return
         yield json.loads(line.decode("utf-8").lstrip("data:").rstrip("/n"))
 
-for chunk in query({
+chunks = query({
     "model": "meta-llama/Llama-3.1-8B-Instruct",
     "messages": [
         {
@@ -23,5 +23,7 @@ def query(payload):
     ],
     "max_tokens": 500,
     "stream": True,
-}):
+})
+
+for chunk in chunks:
     print(chunk["choices"][0]["delta"]["content"], end="")

packages/tasks-gen/snippets-fixtures/conversational-llm-stream/1.requests.together.py

@@ -13,7 +13,7 @@ def query(payload):
             return
         yield json.loads(line.decode("utf-8").lstrip("data:").rstrip("/n"))
 
-for chunk in query({
+chunks = query({
     "model": "<together alias for meta-llama/Llama-3.1-8B-Instruct>",
     "messages": [
         {
@@ -23,5 +23,7 @@ def query(payload):
     ],
     "max_tokens": 500,
     "stream": True,
-}):
+})
+
+for chunk in chunks:
     print(chunk["choices"][0]["delta"]["content"], end="")

packages/tasks-gen/snippets-fixtures/conversational-vlm-stream/1.requests.fireworks-ai.py

@@ -13,7 +13,7 @@ def query(payload):
             return
         yield json.loads(line.decode("utf-8").lstrip("data:").rstrip("/n"))
 
-for chunk in query({
+chunks = query({
     "model": "<fireworks-ai alias for meta-llama/Llama-3.2-11B-Vision-Instruct>",
     "messages": [
         {
@@ -34,5 +34,7 @@ def query(payload):
     ],
     "max_tokens": 500,
     "stream": True,
-}):
+})
+
+for chunk in chunks:
     print(chunk["choices"][0]["delta"]["content"], end="")

packages/tasks-gen/snippets-fixtures/conversational-vlm-stream/1.requests.hf-inference.py

@@ -13,7 +13,7 @@ def query(payload):
             return
         yield json.loads(line.decode("utf-8").lstrip("data:").rstrip("/n"))
 
-for chunk in query({
+chunks = query({
     "model": "meta-llama/Llama-3.2-11B-Vision-Instruct",
     "messages": [
         {
@@ -34,5 +34,7 @@ def query(payload):
     ],
     "max_tokens": 500,
     "stream": True,
-}):
+})
+
+for chunk in chunks:
     print(chunk["choices"][0]["delta"]["content"], end="")