Merge branch 'main' into runners2

.gitignore

@@ -171,3 +171,5 @@ cython_debug/
 
 **/*.xcodeproj/*
 .aider*
+
+exo/tinychat/images/*.png

README.md

@@ -18,6 +18,8 @@ exo: Run your own AI cluster at home with everyday devices. Maintained by [exo l
 [![Tests](https://dl.circleci.com/status-badge/img/circleci/TrkofJDoGzdQAeL6yVHKsg/4i5hJuafuwZYZQxbRAWS71/tree/main.svg?style=svg)](https://dl.circleci.com/status-badge/redirect/circleci/TrkofJDoGzdQAeL6yVHKsg/4i5hJuafuwZYZQxbRAWS71/tree/main)
 [![License: GPL v3](https://img.shields.io/badge/License-GPLv3-blue.svg)](https://www.gnu.org/licenses/gpl-3.0)
 
+<a href="https://trendshift.io/repositories/11849" target="_blank"><img src="https://trendshift.io/api/badge/repositories/11849" alt="exo-explore%2Fexo | Trendshift" style="width: 250px; height: 55px;" width="250" height="55"/></a>
+
 </div>
 
 ---
@@ -38,7 +40,7 @@ We also welcome contributions from the community. We have a list of bounties in
 
 ### Wide Model Support
 
-exo supports different models including LLaMA ([MLX](exo/inference/mlx/models/llama.py) and [tinygrad](exo/inference/tinygrad/models/llama.py)), Mistral, LlaVA, Qwen and Deepseek.
+exo supports different models including LLaMA ([MLX](exo/inference/mlx/models/llama.py) and [tinygrad](exo/inference/tinygrad/models/llama.py)), Mistral, LlaVA, Qwen, and Deepseek.
 
 ### Dynamic Model Partitioning
 
@@ -100,13 +102,13 @@ source install.sh
 
 - There are a number of things users have empirically found to improve performance on Apple Silicon Macs:
 
-1. Upgrade to the latest version of MacOS 15.
+1. Upgrade to the latest version of macOS Sequoia.
 2. Run `./configure_mlx.sh`. This runs commands to optimize GPU memory allocation on Apple Silicon Macs.
 
 
 ## Documentation
 
-### Example Usage on Multiple MacOS Devices
+### Example Usage on Multiple macOS Devices
 
 #### Device 1:
 
@@ -177,9 +179,9 @@ curl http://localhost:52415/v1/chat/completions \
    }'
 ```
 
-### Example Usage on Multiple Heterogenous Devices (MacOS + Linux)
+### Example Usage on Multiple Heterogenous Devices (macOS + Linux)
 
-#### Device 1 (MacOS):
+#### Device 1 (macOS):
 
 ```sh
 exo
@@ -244,7 +246,7 @@ python3 format.py ./exo
 
 ## Known Issues
 
-- On some versions of MacOS/Python, certificates are not installed properly which can lead to SSL errors (e.g. SSL error with huggingface.co). To fix this, run the Install Certificates command, usually:
+- On certain versions of Python on macOS, certificates may not installed correctly, potentially causing SSL errors (e.g., when accessing huggingface.co). To resolve this, run the `Install Certificates` command, typicall as follows:
 
 ```sh
 /Applications/Python 3.x/Install Certificates.command

configure_mlx.sh

@@ -1,4 +1,4 @@
-#!/bin/bash
+#!/usr/bin/env bash
 
 # Get the total memory in MB
 TOTAL_MEM_MB=$(($(sysctl -n hw.memsize) / 1024 / 1024))

examples/function_calling.py

@@ -0,0 +1,111 @@
+import json
+import re
+import requests
+
+def get_current_weather(location: str, unit: str = "celsius"):
+  """Mock weather data function"""
+  # Hardcoded response for demo purposes
+  return {
+    "location": location,
+    "temperature": 22 if unit == "celsius" else 72,
+    "unit": unit,
+    "forecast": "Sunny with light clouds"
+  }
+
+def try_parse_tool_calls(content: str):
+  """Try parse the tool calls."""
+  tool_calls = []
+  offset = 0
+  for i, m in enumerate(re.finditer(r"<tool_call>\n(.+)?\n</tool_call>", content)):
+    if i == 0:
+      offset = m.start()
+    try:
+      func = json.loads(m.group(1))
+      tool_calls.append({"type": "function", "function": func})
+      if isinstance(func["arguments"], str):
+        func["arguments"] = json.loads(func["arguments"])
+    except json.JSONDecodeError as e:
+      print(f"Failed to parse tool calls: the content is {m.group(1)} and {e}")
+      pass
+  if tool_calls:
+    if offset > 0 and content[:offset].strip():
+      c = content[:offset]
+    else:
+      c = ""
+    return {"role": "assistant", "content": c, "tool_calls": tool_calls}
+  return {"role": "assistant", "content": re.sub(r"<\|im_end\|>$", "", content)}
+
+def chat_completion(messages):
+  """Send chat completion request to local server"""
+  response = requests.post(
+    "http://localhost:52415/v1/chat/completions",
+    json={
+      "model": "qwen-2.5-1.5b",
+      "messages": messages,
+      "tools": [{
+        "type": "function",
+        "function": {
+          "name": "get_current_weather",
+          "description": "Get the current weather in a given location",
+          "parameters": {
+            "type": "object",
+            "properties": {
+              "location": {
+                "type": "string",
+                "description": "The city and state, e.g. San Francisco, CA"
+              },
+              "unit": {
+                "type": "string",
+                "enum": ["celsius", "fahrenheit"]
+              }
+            },
+            "required": ["location"]
+          }
+        }
+      }],
+      "tool_choice": "auto"
+    }
+  )
+  return response.json()
+
+def main():
+  # Initial conversation
+  messages = [{
+    "role": "user",
+    "content": "Hi there, what's the weather in Boston?"
+  }]
+
+  # Get initial response
+  response = chat_completion(messages)
+  print(f"First response: {response}")
+  assistant_message = try_parse_tool_calls(response["choices"][0]["message"]["content"])
+  messages.append(assistant_message)
+
+  # If there are tool calls, execute them and continue conversation
+  if "tool_calls" in assistant_message:
+    for tool_call in assistant_message["tool_calls"]:
+      if tool_call["function"]["name"] == "get_current_weather":
+        args = tool_call["function"]["arguments"]
+        weather_data = get_current_weather(**args)
+
+        # Add tool response to messages
+        messages.append({
+          "role": "tool",
+          "content": json.dumps(weather_data),
+          "name": tool_call["function"]["name"]
+        })
+
+    # Get final response with weather data
+    response = chat_completion(messages)
+    print(f"Final response: {response}")
+    messages.append({
+      "role": "assistant",
+      "content": response["choices"][0]["message"]["content"]
+    })
+
+  # Print full conversation
+  for msg in messages:
+    print(f"\n{msg['role'].upper()}: {msg['content']}")
+
+if __name__ == "__main__":
+  main()