[GH-ISSUE #10038] Using Qwen as agent in VS Code #68640

New Issue

GiteaMirror · 2026-05-04T14:41:00-05:00

GiteaMirror commented

2026-05-04 14:41:00 -05:00

Originally created by @Chetosmaister on GitHub (Mar 29, 2025).
Original GitHub issue: https://github.com/ollama/ollama/issues/10038

I am currently using qwen2.5-coder-14b-instruct-q5_k_m in VS Code through Continue and Ollama. It works perfectly in the Continue chat but I am not able to use it as agent, it pops out a message saying there are no tools capabilities.

In this thread: https://github.com/ollama/ollama/issues/8588 rick-github explains: qwen2.5:14b-instruct-q4_K_M support tools capabilities

Do you think:

qwen2.5-coder-14b-instruct-q4_k_m could solve this? Besides being the coder version by reading your response to this thread I was wondering if this could work for me. Thank you so much in advance if you consider to reply me! This is my very first interaction in Github :)

Originally created by @Chetosmaister on GitHub (Mar 29, 2025). Original GitHub issue: https://github.com/ollama/ollama/issues/10038 I am currently using qwen2.5-coder-14b-instruct-q5_k_m in VS Code through Continue and Ollama. It works perfectly in the Continue chat but I am not able to use it as agent, it pops out a message saying there are no tools capabilities. In this thread: https://github.com/ollama/ollama/issues/8588 [rick-github](https://github.com/rick-github) explains: qwen2.5:14b-instruct-q4_K_M support tools capabilities Do you think: qwen2.5-coder-14b-instruct-q4_k_m could solve this? Besides being the coder version by reading your response to this thread I was wondering if this could work for me. Thank you so much in advance if you consider to reply me! This is my very first interaction in Github :)

GiteaMirror added the question label 2026-05-04 14:41:00 -05:00

GiteaMirror closed this issue

2026-05-04 14:41:02 -05:00

GiteaMirror commented

2026-05-04 14:41:04 -05:00

@rick-github commented on GitHub (Mar 29, 2025):

qwen2.5-coder in the ollama library uses tools:

$ ./tool-test.py --model qwen2.5-coder:14b-instruct-q5_k_m --prompt "what operating system are you running on?"
calling get_operating_system({})
The operating system I am running on is Linux, specifically version 6.8.0-47-generic x86_64 with glibc 2.39.

So it's not clear why Continue is reporting no tools capabilities. Did you pull the model from ollama or import it from somewhere else?

@rick-github commented on GitHub (Mar 29, 2025): qwen2.5-coder in the ollama library uses tools: ```console $ ./tool-test.py --model qwen2.5-coder:14b-instruct-q5_k_m --prompt "what operating system are you running on?" calling get_operating_system({}) The operating system I am running on is Linux, specifically version 6.8.0-47-generic x86_64 with glibc 2.39. ``` So it's not clear why Continue is reporting no tools capabilities. Did you pull the model from ollama or import it from somewhere else?

GiteaMirror commented

2026-05-04 14:41:06 -05:00

@Chetosmaister commented on GitHub (Mar 30, 2025):

I pulled it from ollama through git cmd terminal

@Chetosmaister commented on GitHub (Mar 30, 2025): I pulled it from ollama through git cmd terminal

GiteaMirror commented

2026-05-04 14:41:08 -05:00

@rick-github commented on GitHub (Mar 30, 2025):

Do you mean ollama pull qwen2.5-coder:14b-instruct-q5_k_m?

@rick-github commented on GitHub (Mar 30, 2025): Do you mean `ollama pull qwen2.5-coder:14b-instruct-q5_k_m`?

GiteaMirror commented

2026-05-04 14:41:10 -05:00

@Chetosmaister commented on GitHub (Mar 30, 2025):

UPDATE: I use Windows11. q5_k_m version was directly downloaded by the huggingface page. Then I created a modelfile to run the model via ollama to Continue. Chat was good, but Agent was not available. ERROR: No tool capabilities.

I have downloaded q4_k_m version. This time I pulled the model: ollama pull qwen2.5-coder:14b-instruct-q4_k_m but same error would pop up.

PARTIAL SOLUTION:
I disconnected my PC from the router to work offline, this time I was able to select the agent button, but It does not perform as I expected. I will need to try diferent solutions I guess..

CURIOSITY:
Both times happened something weird. While asking "Which model are you?" It would respond: "Hi I am Chat GPT-4 created by OpenAI...." or "Hi I am Claude from Antropic..." This blew my mind! On my terminal would do the same.... But surely has something to do with the interaction through Continue because if I turn off my PC when I start it afterwards if not opening VS Code and opening a terminal it would respond properly: "I am Qwen.. form Alibaba..."

@Chetosmaister commented on GitHub (Mar 30, 2025): UPDATE: I use Windows11. q5_k_m version was directly downloaded by the huggingface page. Then I created a modelfile to run the model via ollama to Continue. Chat was good, but Agent was not available. ERROR: No tool capabilities. I have downloaded q4_k_m version. This time I pulled the model: ollama pull qwen2.5-coder:14b-instruct-q4_k_m but same error would pop up. PARTIAL SOLUTION: I disconnected my PC from the router to work offline, this time I was able to select the agent button, but It does not perform as I expected. I will need to try diferent solutions I guess.. CURIOSITY: Both times happened something weird. While asking "Which model are you?" It would respond: "Hi I am Chat GPT-4 created by OpenAI...." or "Hi I am Claude from Antropic..." This blew my mind! On my terminal would do the same.... But surely has something to do with the interaction through Continue because if I turn off my PC when I start it afterwards if not opening VS Code and opening a terminal it would respond properly: "I am Qwen.. form Alibaba..."

GiteaMirror commented

2026-05-04 14:41:12 -05:00

@rick-github commented on GitHub (Mar 30, 2025):

UPDATE: I use Windows11. q5_k_m version was directly downloaded by the huggingface page. Then I created a modelfile to run the model via ollama to Continue. Chat was good, but Agent was not available. ERROR: No tool capabilities.

Download the model from ollama. ollama pull qwen2.5-coder:14b-instruct-q5_k_m

@rick-github commented on GitHub (Mar 30, 2025): > UPDATE: I use Windows11. q5_k_m version was directly downloaded by the huggingface page. Then I created a modelfile to run the model via ollama to Continue. Chat was good, but Agent was not available. ERROR: No tool capabilities. Download the model from ollama. `ollama pull qwen2.5-coder:14b-instruct-q5_k_m`

GiteaMirror commented

2026-05-04 14:41:16 -05:00

@eugene-kamenev commented on GitHub (Apr 3, 2025):

For any model in Continue config you can override capabilities and specify that it supports tools. Then agent mode becomes available. Just take care about tools/tool calls in ollama model chat template.

@eugene-kamenev commented on GitHub (Apr 3, 2025): For any model in Continue config you can override capabilities and specify that it supports tools. Then agent mode becomes available. Just take care about tools/tool calls in ollama model chat template.

GiteaMirror commented

2026-05-04 14:41:22 -05:00

@Chetosmaister commented on GitHub (Apr 4, 2025):

For any model in Continue config you can override capabilities and specify that it supports tools. Then agent mode becomes available. Just take care about tools/tool calls in ollama model chat template.

Thank you very much Eugene, I´ll investigate. MAIN GOAL WHEN I OPENED THIS THREAD was being able to use QWEN2.5 as an agent and thanks to both rick and you I think I´ve got that covered. THANK YOU BOTH VERY MUCH!!!! Besides that, I am new into development and want to optimize my hardware: ProArt X670 / 7900x / Rtx 4080 Super OC / 64 GB RAM / Windows 11. I have been playing around with different IDE's (Cursor, Trae, Replit...) and figured out I might end up with a similar experience in VS Code for free, my main objective is developing modern and scalable Next.js based apps. I don't want to sound delusional, but I think it might work for me adding different extensions to enhance my VS Code experience so it can help me through the process. I have selected these extensions: 1. "Live Preview (Microsoft)" to replace the Replit "Realtme Preview" feature, 2. "Continue" to replace the agent feature of other IDE's using QWEN2.5 locally, 3. "ESLint (Microsoft)" to maintain code consistency and prevent common bugs and 4. "Prettier - Code Formatter" to maintain style consistency and automatically code formatting... I will have to study how to set up properly all extensions so they can work seamlessly together. I feel a proper set up prior to start a project is crucial for me. If you feel like giving me any advice I would really appreciate it. I do apologize in advance if I shouldn't interact like this in my thread and do also apologize for my English as I do not use any translation tool and I am from Spain! :)

@Chetosmaister commented on GitHub (Apr 4, 2025): > For any model in Continue config you can override capabilities and specify that it supports tools. Then agent mode becomes available. Just take care about tools/tool calls in ollama model chat template. Thank you very much Eugene, I´ll investigate. MAIN GOAL WHEN I OPENED THIS THREAD was being able to use QWEN2.5 as an agent and thanks to both rick and you I think I´ve got that covered. THANK YOU BOTH VERY MUCH!!!! Besides that, I am new into development and want to optimize my hardware: ProArt X670 / 7900x / Rtx 4080 Super OC / 64 GB RAM / Windows 11. I have been playing around with different IDE's (Cursor, Trae, Replit...) and figured out I might end up with a similar experience in VS Code for free, my main objective is developing modern and scalable Next.js based apps. I don't want to sound delusional, but I think it might work for me adding different extensions to enhance my VS Code experience so it can help me through the process. I have selected these extensions: 1. "Live Preview (Microsoft)" to replace the Replit "Realtme Preview" feature, 2. "Continue" to replace the agent feature of other IDE's using QWEN2.5 locally, 3. "ESLint (Microsoft)" to maintain code consistency and prevent common bugs and 4. "Prettier - Code Formatter" to maintain style consistency and automatically code formatting... I will have to study how to set up properly all extensions so they can work seamlessly together. I feel a proper set up prior to start a project is crucial for me. If you feel like giving me any advice I would really appreciate it. I do apologize in advance if I shouldn't interact like this in my thread and do also apologize for my English as I do not use any translation tool and I am from Spain! :)

GiteaMirror commented

2026-05-04 14:41:26 -05:00

@esayli commented on GitHub (Apr 22, 2025):

@Chetosmaister I also tried qwen2.5-coder:14b-instruct-q5_k_m with Vscode insders latest , and ollama. But when I ask to edit a file in agent mode, it just gives me the instruction to edit that myself. It does not use tools. Yours editing files ?

@esayli commented on GitHub (Apr 22, 2025): @Chetosmaister I also tried qwen2.5-coder:14b-instruct-q5_k_m with Vscode insders latest , and ollama. But when I ask to edit a file in agent mode, it just gives me the instruction to edit that myself. It does not use tools. Yours editing files ?

GiteaMirror commented

2026-05-04 14:41:30 -05:00

@LFd3v commented on GitHub (Apr 22, 2025):

$ ./tool-test.py

@rick-github Sorry to disturb you (trying to create minimum disturbance I am using a recent closed thread), but could you please share your tool-test.py, maybe in a gist? There is a tools.py in ollama-python repo but your seems to be more complete. Thank you in advance, and also for all the work in Ollama. Regards.

@LFd3v commented on GitHub (Apr 22, 2025): > $ ./tool-test.py @rick-github Sorry to disturb you (trying to create minimum disturbance I am using a recent closed thread), but could you please share your tool-test.py, maybe in a gist? There is a tools.py in ollama-python repo but your seems to be more complete. Thank you in advance, and also for all the work in Ollama. Regards.

GiteaMirror commented

2026-05-04 14:41:31 -05:00

@rick-github commented on GitHub (Apr 24, 2025):

@LFd3v It's pretty rough but you're welcome to poke around in it.

#!/usr/bin/env python3

import ollama
import argparse
import json
import sys
import readline
import dotenv
import os
import requests

parser = argparse.ArgumentParser()
parser.add_argument("-m", "--model", help="Set model to use", default="qwen2.5:7b")
parser.add_argument("-s", "--system", help="Set system prompt")
parser.add_argument("-c", "--context", help="Set num_ctx", type=int, default=20480)
parser.add_argument("-G", "--gpu", help="Set num_gpu", type=int, default=-1)
parser.add_argument("-T", "--temperature", help="Set temperature", type=int, default=1)
parser.add_argument("-t", "--tools", help="Set tools to use.  Use `list` to see available tools", default="add,get_operating_system")
parser.add_argument("-b", "--sandbox", help="Make the run_python_script tool use a sandbox", default=False, action="store_true")
parser.add_argument("-v", "--verbose", help="Show what's going on", action='count', default=1)
parser.add_argument("-q", "--quiet", help="Turn off debugging statements", dest='verbose', action='store_const', const=0)
parser.add_argument("-p", "--prompts", help="One or prompts", action='append', default=None)
parser.add_argument("-i", "--interactive", help="Enter interactive mode", action='store_true', default=False)
parser.add_argument("--test", help="Run canned prompts", action='store_true', default=False)
args = parser.parse_args()

all_tools = []

def tool(f):
    all_tools.append(f)
    return f

@tool
def add(l: list) -> float:
    """Adds 2 or more numbers and returns the result"""
    return sum(l)

@tool
def multiply(x: float, y: float) -> float:
    """Multiplies 2 numbers and returns the result"""
    return x * y

@tool
def subtract(x: float, y: float) -> float:
    """Substracts y from x and returns the result"""
    return x - y

@tool
def divide(x: float, y: float) -> float:
    """Divide x by y and returns the result"""
    return x / y

@tool
def count(l: list) -> int:
    """Counts the number of elements in a list and returns the result"""
    return len(l)

@tool
def power(x: float, y: float) -> float:
    """Raises x to the y power and returns the result"""
    return float(x) ** float(y)

@tool
def get_weather(country: str, city: str, unit: str) -> str:
    """ Returns the weather in the requested location """
    from bs4 import BeautifulSoup
    response = requests.get(f'https://www.timeanddate.com/weather/{country.replace(" ","-")}/{city.replace(" ","-")}')
    soup = BeautifulSoup(response.text, 'html.parser')
    r = []
    for paragraph in soup.find_all('p'):
      r.append(paragraph.text)
    return "\n".join(r)

@tool
def get_datetime() -> str:
    """
    Returns the current date and time as JSON data structure, in the format
      '{{"fulldate":"<fulldate>","date":"<date>","time":"<time>"}}'
    Time is in 24 hour format with the day starting at 00:00.
    """
    import datetime
    data = {
      "fulldate": datetime.datetime.now().strftime("%A, %B %d, %Y %H:%M %Z"),
      "date": datetime.datetime.now().strftime("%A, %B %d, %Y"),
      "time": datetime.datetime.now().strftime("%H:%M")
    }
    return json.dumps(data)

@tool
def get_operating_system() -> str:
    """
    Returns a string representing the operating system that
    the model is running on.
    """
    import platform
    return platform.platform()

@tool
def run_shell_command(command: str) -> str:
    """
    Execute a shell command and return its output as a string.
    """
    import subprocess

    try:
        # shell=True allows shell-specific features like pipes and redirections
        result = subprocess.run(
            command,
            shell=True,
            check=True,  # Raises CalledProcessError if return code is non-zero
            text=True,   # Return string instead of bytes
            capture_output=True  # Capture stdout and stderr
        )
        return result.stdout.strip()
    except subprocess.CalledProcessError as e:
        # Include error message in the exception
        raise subprocess.CalledProcessError(
            e.returncode,
            e.cmd,
            e.output,
            f"Command failed with exit code {e.returncode}: {e.stderr}"
        )

# requires searxng instance at :5342
@tool
def web_search(query:str) -> str:
    """
    Searches the web using the supplied query and returns
    a JSON structure with a summary of the results.
    If the response from this tool doesn't provide the needed
    information, ignore it and inform the client.
    """
    import multiprocessing
    from collections import deque
    from bs4 import BeautifulSoup

    r = []
    response = requests.get("http://localhost:5342/search", params={"q":query,"format":"json"})
    if response.status_code != requests.codes.ok:
      return response
    results = response.json().get("results", [])
    if len(results) < 1:
      return "no results"
#    if x := response.json().get("infoboxes"):
#      r.append(x)

    taskq = multiprocessing.Queue()
    resultq = multiprocessing.Queue()

    def worker(id, taskq, resultq):
      while True:
        try:
          url = taskq.get(timeout=1.0)
          if url is None:
            break
          result = {"url":url,"status":0,"response":""}
          try:
            response = requests.get(url, timeout=30)
            if response.status_code == requests.codes.ok:
              s = []
              soup = BeautifulSoup(response.content, 'html.parser')
              for paragraph in soup.find_all('p'):
                s.append(paragraph.text)
              result["response"] = "\n".join(s)
              result["status"] = requests.codes.ok
          except Exception as e:
            result["response"] = e
          resultq.put(result)
        except multiprocessing.queues.Empty:
          continue
        except Exception as e:
          break

    workers = []
    for i in range(min(20, len(results))):
      p = multiprocessing.Process(
          target=worker,
          args=(i, taskq, resultq),
          name=f"worker-{i}"
        )
      p.start()
      workers.append(p)

    urls = deque([result["url"] for result in results])
    tasks = len(results)
    for i in range(len(workers)):
      if urls:
        taskq.put(urls.popleft())
    while tasks:
      tasks -= 1
      result = resultq.get()
      if result["status"] == requests.codes.ok and len(result["response"]) > 10:
        r.append(result)
      if len(r) == 10:
        break
      if urls:
        taskq.put(urls.popleft())
    for _, p in enumerate(workers):
      p.terminate()
      p.join()
    return(r)

@tool
def read_file(filename:str) -> str:
    """
    Read the local file `filename` and return the contents.
    """
    with open(filename, "r") as f:
      return f.read()
    return ""

@tool
def http_get(url:str) -> str:
    """
    Retrieves a specific page from the internet.
    """
    return requests.get(url, timeout=10).text

@tool
def http_post(url:str, form:str) -> str:
    """
    Post a form to the specified url.
    """
    result = requests.post(url, data=form, timeout=10)
    if result.status_code == requests.codes.ok:
      return "success"
    return "fail"

@tool
def describe_contents_of_image_file(filename:str, query:str = None, model:str = None) -> str:
    """
    The file referenced by the `filename` argument is passed to
    a model which reads the contents of the file and returns
    a description of the image.

    Args:
      filename: name of the file to examine
      model: optional LLM
      query: optional query
    """
    f = filename
    if not (filename.startswith("/") or filename.startswith("./")):
      f = "./" + filename
    if not os.path.isfile(f):
      raise Exception(f"file '{filename}' does not exist")
    if not model:
      model = "gemma3:4b"
    if not query:
      query = 'describe this image'
    response = ollama.chat(
      model=model,
      messages=[{
        'role': 'user',
        'content': query,
        'images': [f]
      }]
    )
    return {"filename":filename, "description": response.message.content}

@tool
def get_current_directory() -> str:
    """
    Return the current working directory.
    """
    return run_shell_command("pwd")

@tool
def list_directory(directory: str) -> str:
    """
    Return the list of files in the named directory.
    """
    if directory != '~':
      directory = f"'{directory}'"
    return run_shell_command(f"ls {directory}")

@tool
def run_python_script(script: str) -> str:
    """
    Takes a python script as an argument, executes the script, and returns the results.
    The script must print results to stdout or save them in a file.
    If the 'run_python_script' tool returns no response, try again and make the script explicitly save the graph to a file.
    If the script requires modules, they should be added with "!pip install" commands.
    """
    if args.sandbox:
      return run_python_script_sandbox(script)
    return run_python_script_nosandbox(script)

@tool
def run_python_script_nosandbox(script: str) -> str:
    """
    Takes a python script as an argument, executes the script, and returns the results
    """
    from io import StringIO
    from contextlib import redirect_stdout

    f = StringIO()
    with redirect_stdout(f):
      try:
        ns = {}
        exec(script, globals(), ns)
        result = f.getvalue()
        if not result:
          result = ns
      except Exception as e:
        result = str(e)
    return result

@tool
def run_python_script_sandbox(script: str) -> str:
    """
    Takes a python script as an argument, executes the script, and returns the results
    """
# pip install epicbox
# d=$(mktemp -d) && printf "FROM python:3.8-bookworm\nRUN pip install yfinance matplotlib\n" > $d/Dockerfile && docker build -f $d/Dockerfile -t python-sandbox $d && rm $d/Dockerfile && rmdir $d
    import epicbox
    import tempfile
    import base64

    epicbox.configure(
        profiles=[
            epicbox.Profile('python', 'python-sandbox', network_disabled=False)
        ]
    )
    main = []
    scr = []
    timeout = 60
    for line in script.split("\n"):
      if line.startswith("# pip install"):
        timeout = timeout + 60
        main.append("pip install -qqq " + line[13:])
        continue
      if line.startswith("!pip install"):
        timeout = timeout + 60
        main.append("pip install -qqq " + line[12:])
        continue
      scr.append(line)
    main.append("""python3 ./script.py && { files=$(ls|egrep -v "^(main.sh|script.py)") ; [ -n "$files" ] && { comma="" ; echo '{"files": [' ; for i in $files ; do echo $comma'{ "name": "'$i'", "content": "'$(base64 -w0 "$i")'" }' ; comma=, ; done ; echo ']}' ; } ; : ; }""")
    files = [
          { "name":"main.sh", "content": bytes("\n".join(main).encode()) },
          { "name":"script.py", "content": bytes("\n".join(scr).encode()) },
    ]
    for f in files:
      print(f["name"])
      print(f["content"])
    result = epicbox.run("python", "sh main.sh", files=files, limits={"realtime":timeout,"cputime":timeout,"memory":1000})
    print(result)
    if result["timeout"]:
      return "script failed to start"
    if result["exit_code"] == 0:
      if len(result["stdout"]) == 0:
        return "There was no output from the script."
      try:
        j = json.loads(result["stdout"])
        res = []
        for f in j["files"]:
          if "name" in f and "content" in f:
            try:
              content = base64.b64decode(f["content"])
            except Exception as e:
              content = f["content"]
            with tempfile.NamedTemporaryFile(delete=False) as temp:
              temp.write(content)
              if f["name"].endswith(tuple([".jpg", ".jpeg", ".png", ".gif"])):
                res.append(f'![{f["name"]}]({temp.name})')
              else:
                res.append(f'[{f["name"]}]({temp.name})')
        return "\n".join(res)
      except Exception as e:
        print(f"failed to load json: {e}")
        pass
      return result["stdout"]
    return result["stderr"]

@tool
def send_email(recipient:str, message:str, sender:str = None):
  """
  Sends an email message to the designated recipient.  The `sender`
  argument is optional and will be inferred from the auth details
  if not supplied.
  """
  import smtplib, ssl
  with smtplib.SMTP_SSL(os.getenv("SMTP_GATEWAY", "smtp.gmail.com"), 465, context=ssl.create_default_context()) as server:
    username = os.getenv("SMTP_USERNAME")
    password = os.getenv("SMTP_PASSWORD")
    if username and password:
        server.login(username, password)
    return server.sendmail(sender or username, recipient, message)

@tool
def post_to_bluesky(message:str):
  """
  Make a post to the bsky social network.
  """
  import atproto
  provider = os.getenv("BSKY_PROVIDER", "bsky.social")
  username = os.getenv("BSKY_USERNAME")
  password = os.getenv("BSKY_PASSWORD")
  if "." not in username:
    username = f"{username}.{provider}"
  client = atproto.Client()
  client.login(username, password)
  return client.send_post(message)

@tool
def get_recent_mastodon_posts(count:int = 5):
  """
  Get recent mastodon posts.
  """
  server = os.getenv("MASTODON_SERVER", "mastodon.social")
  return requests.get(f"https://{server}/api/v1/timelines/public?limit={count}").json()

@tool
def get_ip_address():
  """
  Returns the public IP address of this machine.
  """
  return requests.get("https://icanhazip.com").text

@tool
def get_location():
  """
  Returns the approximate location based on IP address.
  """
  return requests.get("https://mylocation.org", headers={"User-agent": "mozilla"}).text

all_tools_map = {t.__name__:t for t in all_tools}
tools = []

for tool in args.tools.split(","):
  match tool:
    case "all":
      tools = all_tools
    case "none":
      tools = []
    case "list":
      print("\n".join([f.__name__ for f in all_tools]))
      sys.exit(0)
    case _:
      if f := all_tools_map.get(tool):
        tools.append(f)
      
tools_map = {f.__name__:f for f in tools}

options = {
  "temperature":args.temperature,
  "num_ctx":args.context,
  "num_gpu":args.gpu,
}

def chat(messages, prompt):
  messages.append({"role":"user", "content":prompt})
  for _ in range(10):
    print("== " + json.dumps(messages, indent=4, default=str) + " ==>LLM") if args.verbose > 1 else True
    response = ollama.chat(model=args.model, options=options, messages=messages, tools=tools)
    if response.message.content.startswith('{"'):
      messages.append({"role":"tool","content":f"Badly formed input {response.message.content}"})
      continue
    messages.append({"role":"assistant","content":response.message.content,"tool_calls":response.message.tool_calls})
    if not response.message.tool_calls:
      break
    for tool in response.message.tool_calls:
        if tool_func := tools_map.get(tool.function.name):
          print(f"calling {tool.function.name}({tool.function.arguments})") if args.verbose > 0 else True
          try:
            tool_output = tool_func(**tool.function.arguments)
          except Exception as e:
            print(e) if args.verbose > 1 else True
            tool_output = e
          if type(tool_output) == bytes:
            tool_output = tool_output.decode()
          tool_output = str(tool_output)
          messages.append({"role":"tool","content":tool_output})
  print(f"{response.message.content}")
  return messages

dotenv.load_dotenv()

userprompt = ">>> " if sys.stdin.isatty() else ""

messages = [{"role":"system", "content":args.system}] if args.system else []

if args.test:
  args.prompts = [
    "what time is it?",
    "what is 10 ^ 1.34?",
    "what files are in the /tmp directory?",
    "what operating system are you running on?",
    "what is my ip address?",
    "what is my location?",
    "What's the weather where I am?",
    "What's the weather in Paris?",
    "count the number of files in the current directory",
    "when is sunrise in New Zealand tomorrow?",
    "summarize the contents of the file /etc/os-release",
    "what happened this week in AI?",
    "describe what's in ./filename.png",
    "What are the 3 most recent posts in mastodon?",
    "plot x^2",
  ]

if args.prompts:
  for p in args.prompts:
    print(f">>> {p}") if args.test else True
    messages = chat(messages, p)
if args.interactive or not args.prompts:
  while True:
    try:
      prompt = input(userprompt)
    except:
      print()
      break
    if prompt.startswith('"""'):
      long_prompt = prompt
      while True:
        try:
          prompt = input("... ")
        except:
          print()
          break
        if prompt.startswith('"""'):
          break
        long_prompt += prompt
      prompt = long_prompt
    if prompt.startswith('!'):
      os.system(prompt[1:])
      continue
    if prompt == "/bye":
      break
    if prompt == "/clear":
      messages = [{"role":"system", "content":args.system}] if args.system else []
      continue
    if prompt == "/dump":
      print(json.dumps(messages, indent=4, default=str))
      continue
    if prompt.startswith("/set"):
      cmd = prompt[5:].split()
      if cmd[0] == "parameter":
        options[cmd[1]] = float(cmd[2])
      if cmd[0] == "system":
        system = prompt.split(maxsplit=2)[2]
        messages.append({"role":"system", "content":system})
      continue
    messages = chat(messages, prompt)

@rick-github commented on GitHub (Apr 24, 2025): @LFd3v It's pretty rough but you're welcome to poke around in it. ```python #!/usr/bin/env python3 import ollama import argparse import json import sys import readline import dotenv import os import requests parser = argparse.ArgumentParser() parser.add_argument("-m", "--model", help="Set model to use", default="qwen2.5:7b") parser.add_argument("-s", "--system", help="Set system prompt") parser.add_argument("-c", "--context", help="Set num_ctx", type=int, default=20480) parser.add_argument("-G", "--gpu", help="Set num_gpu", type=int, default=-1) parser.add_argument("-T", "--temperature", help="Set temperature", type=int, default=1) parser.add_argument("-t", "--tools", help="Set tools to use. Use `list` to see available tools", default="add,get_operating_system") parser.add_argument("-b", "--sandbox", help="Make the run_python_script tool use a sandbox", default=False, action="store_true") parser.add_argument("-v", "--verbose", help="Show what's going on", action='count', default=1) parser.add_argument("-q", "--quiet", help="Turn off debugging statements", dest='verbose', action='store_const', const=0) parser.add_argument("-p", "--prompts", help="One or prompts", action='append', default=None) parser.add_argument("-i", "--interactive", help="Enter interactive mode", action='store_true', default=False) parser.add_argument("--test", help="Run canned prompts", action='store_true', default=False) args = parser.parse_args() all_tools = [] def tool(f): all_tools.append(f) return f @tool def add(l: list) -> float: """Adds 2 or more numbers and returns the result""" return sum(l) @tool def multiply(x: float, y: float) -> float: """Multiplies 2 numbers and returns the result""" return x * y @tool def subtract(x: float, y: float) -> float: """Substracts y from x and returns the result""" return x - y @tool def divide(x: float, y: float) -> float: """Divide x by y and returns the result""" return x / y @tool def count(l: list) -> int: """Counts the number of elements in a list and returns the result""" return len(l) @tool def power(x: float, y: float) -> float: """Raises x to the y power and returns the result""" return float(x) ** float(y) @tool def get_weather(country: str, city: str, unit: str) -> str: """ Returns the weather in the requested location """ from bs4 import BeautifulSoup response = requests.get(f'https://www.timeanddate.com/weather/{country.replace(" ","-")}/{city.replace(" ","-")}') soup = BeautifulSoup(response.text, 'html.parser') r = [] for paragraph in soup.find_all('p'): r.append(paragraph.text) return "\n".join(r) @tool def get_datetime() -> str: """ Returns the current date and time as JSON data structure, in the format '{{"fulldate":"<fulldate>","date":"<date>","time":"<time>"}}' Time is in 24 hour format with the day starting at 00:00. """ import datetime data = { "fulldate": datetime.datetime.now().strftime("%A, %B %d, %Y %H:%M %Z"), "date": datetime.datetime.now().strftime("%A, %B %d, %Y"), "time": datetime.datetime.now().strftime("%H:%M") } return json.dumps(data) @tool def get_operating_system() -> str: """ Returns a string representing the operating system that the model is running on. """ import platform return platform.platform() @tool def run_shell_command(command: str) -> str: """ Execute a shell command and return its output as a string. """ import subprocess try: # shell=True allows shell-specific features like pipes and redirections result = subprocess.run( command, shell=True, check=True, # Raises CalledProcessError if return code is non-zero text=True, # Return string instead of bytes capture_output=True # Capture stdout and stderr ) return result.stdout.strip() except subprocess.CalledProcessError as e: # Include error message in the exception raise subprocess.CalledProcessError( e.returncode, e.cmd, e.output, f"Command failed with exit code {e.returncode}: {e.stderr}" ) # requires searxng instance at :5342 @tool def web_search(query:str) -> str: """ Searches the web using the supplied query and returns a JSON structure with a summary of the results. If the response from this tool doesn't provide the needed information, ignore it and inform the client. """ import multiprocessing from collections import deque from bs4 import BeautifulSoup r = [] response = requests.get("http://localhost:5342/search", params={"q":query,"format":"json"}) if response.status_code != requests.codes.ok: return response results = response.json().get("results", []) if len(results) < 1: return "no results" # if x := response.json().get("infoboxes"): # r.append(x) taskq = multiprocessing.Queue() resultq = multiprocessing.Queue() def worker(id, taskq, resultq): while True: try: url = taskq.get(timeout=1.0) if url is None: break result = {"url":url,"status":0,"response":""} try: response = requests.get(url, timeout=30) if response.status_code == requests.codes.ok: s = [] soup = BeautifulSoup(response.content, 'html.parser') for paragraph in soup.find_all('p'): s.append(paragraph.text) result["response"] = "\n".join(s) result["status"] = requests.codes.ok except Exception as e: result["response"] = e resultq.put(result) except multiprocessing.queues.Empty: continue except Exception as e: break workers = [] for i in range(min(20, len(results))): p = multiprocessing.Process( target=worker, args=(i, taskq, resultq), name=f"worker-{i}" ) p.start() workers.append(p) urls = deque([result["url"] for result in results]) tasks = len(results) for i in range(len(workers)): if urls: taskq.put(urls.popleft()) while tasks: tasks -= 1 result = resultq.get() if result["status"] == requests.codes.ok and len(result["response"]) > 10: r.append(result) if len(r) == 10: break if urls: taskq.put(urls.popleft()) for _, p in enumerate(workers): p.terminate() p.join() return(r) @tool def read_file(filename:str) -> str: """ Read the local file `filename` and return the contents. """ with open(filename, "r") as f: return f.read() return "" @tool def http_get(url:str) -> str: """ Retrieves a specific page from the internet. """ return requests.get(url, timeout=10).text @tool def http_post(url:str, form:str) -> str: """ Post a form to the specified url. """ result = requests.post(url, data=form, timeout=10) if result.status_code == requests.codes.ok: return "success" return "fail" @tool def describe_contents_of_image_file(filename:str, query:str = None, model:str = None) -> str: """ The file referenced by the `filename` argument is passed to a model which reads the contents of the file and returns a description of the image. Args: filename: name of the file to examine model: optional LLM query: optional query """ f = filename if not (filename.startswith("/") or filename.startswith("./")): f = "./" + filename if not os.path.isfile(f): raise Exception(f"file '{filename}' does not exist") if not model: model = "gemma3:4b" if not query: query = 'describe this image' response = ollama.chat( model=model, messages=[{ 'role': 'user', 'content': query, 'images': [f] }] ) return {"filename":filename, "description": response.message.content} @tool def get_current_directory() -> str: """ Return the current working directory. """ return run_shell_command("pwd") @tool def list_directory(directory: str) -> str: """ Return the list of files in the named directory. """ if directory != '~': directory = f"'{directory}'" return run_shell_command(f"ls {directory}") @tool def run_python_script(script: str) -> str: """ Takes a python script as an argument, executes the script, and returns the results. The script must print results to stdout or save them in a file. If the 'run_python_script' tool returns no response, try again and make the script explicitly save the graph to a file. If the script requires modules, they should be added with "!pip install" commands. """ if args.sandbox: return run_python_script_sandbox(script) return run_python_script_nosandbox(script) @tool def run_python_script_nosandbox(script: str) -> str: """ Takes a python script as an argument, executes the script, and returns the results """ from io import StringIO from contextlib import redirect_stdout f = StringIO() with redirect_stdout(f): try: ns = {} exec(script, globals(), ns) result = f.getvalue() if not result: result = ns except Exception as e: result = str(e) return result @tool def run_python_script_sandbox(script: str) -> str: """ Takes a python script as an argument, executes the script, and returns the results """ # pip install epicbox # d=$(mktemp -d) && printf "FROM python:3.8-bookworm\nRUN pip install yfinance matplotlib\n" > $d/Dockerfile && docker build -f $d/Dockerfile -t python-sandbox $d && rm $d/Dockerfile && rmdir $d import epicbox import tempfile import base64 epicbox.configure( profiles=[ epicbox.Profile('python', 'python-sandbox', network_disabled=False) ] ) main = [] scr = [] timeout = 60 for line in script.split("\n"): if line.startswith("# pip install"): timeout = timeout + 60 main.append("pip install -qqq " + line[13:]) continue if line.startswith("!pip install"): timeout = timeout + 60 main.append("pip install -qqq " + line[12:]) continue scr.append(line) main.append("""python3 ./script.py && { files=$(ls|egrep -v "^(main.sh|script.py)") ; [ -n "$files" ] && { comma="" ; echo '{"files": [' ; for i in $files ; do echo $comma'{ "name": "'$i'", "content": "'$(base64 -w0 "$i")'" }' ; comma=, ; done ; echo ']}' ; } ; : ; }""") files = [ { "name":"main.sh", "content": bytes("\n".join(main).encode()) }, { "name":"script.py", "content": bytes("\n".join(scr).encode()) }, ] for f in files: print(f["name"]) print(f["content"]) result = epicbox.run("python", "sh main.sh", files=files, limits={"realtime":timeout,"cputime":timeout,"memory":1000}) print(result) if result["timeout"]: return "script failed to start" if result["exit_code"] == 0: if len(result["stdout"]) == 0: return "There was no output from the script." try: j = json.loads(result["stdout"]) res = [] for f in j["files"]: if "name" in f and "content" in f: try: content = base64.b64decode(f["content"]) except Exception as e: content = f["content"] with tempfile.NamedTemporaryFile(delete=False) as temp: temp.write(content) if f["name"].endswith(tuple([".jpg", ".jpeg", ".png", ".gif"])): res.append(f'![{f["name"]}]({temp.name})') else: res.append(f'[{f["name"]}]({temp.name})') return "\n".join(res) except Exception as e: print(f"failed to load json: {e}") pass return result["stdout"] return result["stderr"] @tool def send_email(recipient:str, message:str, sender:str = None): """ Sends an email message to the designated recipient. The `sender` argument is optional and will be inferred from the auth details if not supplied. """ import smtplib, ssl with smtplib.SMTP_SSL(os.getenv("SMTP_GATEWAY", "smtp.gmail.com"), 465, context=ssl.create_default_context()) as server: username = os.getenv("SMTP_USERNAME") password = os.getenv("SMTP_PASSWORD") if username and password: server.login(username, password) return server.sendmail(sender or username, recipient, message) @tool def post_to_bluesky(message:str): """ Make a post to the bsky social network. """ import atproto provider = os.getenv("BSKY_PROVIDER", "bsky.social") username = os.getenv("BSKY_USERNAME") password = os.getenv("BSKY_PASSWORD") if "." not in username: username = f"{username}.{provider}" client = atproto.Client() client.login(username, password) return client.send_post(message) @tool def get_recent_mastodon_posts(count:int = 5): """ Get recent mastodon posts. """ server = os.getenv("MASTODON_SERVER", "mastodon.social") return requests.get(f"https://{server}/api/v1/timelines/public?limit={count}").json() @tool def get_ip_address(): """ Returns the public IP address of this machine. """ return requests.get("https://icanhazip.com").text @tool def get_location(): """ Returns the approximate location based on IP address. """ return requests.get("https://mylocation.org", headers={"User-agent": "mozilla"}).text all_tools_map = {t.__name__:t for t in all_tools} tools = [] for tool in args.tools.split(","): match tool: case "all": tools = all_tools case "none": tools = [] case "list": print("\n".join([f.__name__ for f in all_tools])) sys.exit(0) case _: if f := all_tools_map.get(tool): tools.append(f) tools_map = {f.__name__:f for f in tools} options = { "temperature":args.temperature, "num_ctx":args.context, "num_gpu":args.gpu, } def chat(messages, prompt): messages.append({"role":"user", "content":prompt}) for _ in range(10): print("== " + json.dumps(messages, indent=4, default=str) + " ==>LLM") if args.verbose > 1 else True response = ollama.chat(model=args.model, options=options, messages=messages, tools=tools) if response.message.content.startswith('{"'): messages.append({"role":"tool","content":f"Badly formed input {response.message.content}"}) continue messages.append({"role":"assistant","content":response.message.content,"tool_calls":response.message.tool_calls}) if not response.message.tool_calls: break for tool in response.message.tool_calls: if tool_func := tools_map.get(tool.function.name): print(f"calling {tool.function.name}({tool.function.arguments})") if args.verbose > 0 else True try: tool_output = tool_func(**tool.function.arguments) except Exception as e: print(e) if args.verbose > 1 else True tool_output = e if type(tool_output) == bytes: tool_output = tool_output.decode() tool_output = str(tool_output) messages.append({"role":"tool","content":tool_output}) print(f"{response.message.content}") return messages dotenv.load_dotenv() userprompt = ">>> " if sys.stdin.isatty() else "" messages = [{"role":"system", "content":args.system}] if args.system else [] if args.test: args.prompts = [ "what time is it?", "what is 10 ^ 1.34?", "what files are in the /tmp directory?", "what operating system are you running on?", "what is my ip address?", "what is my location?", "What's the weather where I am?", "What's the weather in Paris?", "count the number of files in the current directory", "when is sunrise in New Zealand tomorrow?", "summarize the contents of the file /etc/os-release", "what happened this week in AI?", "describe what's in ./filename.png", "What are the 3 most recent posts in mastodon?", "plot x^2", ] if args.prompts: for p in args.prompts: print(f">>> {p}") if args.test else True messages = chat(messages, p) if args.interactive or not args.prompts: while True: try: prompt = input(userprompt) except: print() break if prompt.startswith('"""'): long_prompt = prompt while True: try: prompt = input("... ") except: print() break if prompt.startswith('"""'): break long_prompt += prompt prompt = long_prompt if prompt.startswith('!'): os.system(prompt[1:]) continue if prompt == "/bye": break if prompt == "/clear": messages = [{"role":"system", "content":args.system}] if args.system else [] continue if prompt == "/dump": print(json.dumps(messages, indent=4, default=str)) continue if prompt.startswith("/set"): cmd = prompt[5:].split() if cmd[0] == "parameter": options[cmd[1]] = float(cmd[2]) if cmd[0] == "system": system = prompt.split(maxsplit=2)[2] messages.append({"role":"system", "content":system}) continue messages = chat(messages, prompt) ```

Sign in to join this conversation.

Branches Tags

main

dhiltgen/ci

dhiltgen/llama-runner

hoyyeva/anthropic-local-image-path

hoyyeva/anthropic-reference-images-path

parth-anthropic-reference-images-path

brucemacd/download-before-remove

hoyyeva/editor-config-repair

parth-mlx-decode-checkpoints

parth-launch-codex-app

hoyyeva/fix-codex-model-metadata-warning

hoyyeva/qwen

parth/hide-claude-desktop-till-release

hoyyeva/opencode-image-modality

parth-add-claude-code-autoinstall

release_v0.22.0

pdevine/manifest-list

codex/fix-codex-model-metadata-warning

pdevine/addressable-manifest

brucemacd/launch-fetch-reccomended

jmorganca/llama-compat

launch-copilot-cli

hoyyeva/opencode-thinking

release_v0.20.7

parth-auto-save-backup

parth-test

jmorganca/gemma4-audio-replacements

fix-manifest-digest-on-pull

hoyyeva/vscode-improve

brucemacd/install-server-wait

parth/update-claude-docs

brucemac/start-ap-install

pdevine/mlx-update

pdevine/qwen35_vision

drifkin/api-show-fallback

mintlify/image-generation-1773352582

hoyyeva/server-context-length-local-config

jmorganca/faster-reptition-penalties

jmorganca/convert-nemotron

parth-pi-thinking

pdevine/sampling-penalties

jmorganca/fix-create-quantization-memory

dongchen/resumable_transfer_fix

pdevine/sampling-cache-error

jessegross/mlx-usage

hoyyeva/openclaw-config

hoyyeva/app-html

pdevine/qwen3next

brucemacd/sign-sh-install

brucemacd/tui-update

brucemacd/usage-api

jmorganca/launch-empty

fix-app-dist-embed

mxyng/mlx-compile

mxyng/mlx-quant

mxyng/mlx-glm4.7

mxyng/mlx

brucemacd/simplify-model-picker

jmorganca/qwen3-concurrent

fix-glm-4.7-flash-mla-config

drifkin/qwen3-coder-opening-tag

brucemacd/usage-cli

fix-cuda12-fattn-shmem

ollama-imagegen-docs

parth/fix-multiline-inputs

brucemacd/config-docs

mxyng/model-files

mxyng/simple-execute

fix-imagegen-ollama-models

mxyng/async-upload

jmorganca/lazy-no-dtype-changes

imagegen-auto-detect-create

parth/decrease-concurrent-download-hf

fix-mlx-quantize-init

jmorganca/x-cleanup

usage

imagegen-readme

jmorganca/glm-image

mlx-gpu-cd

jmorganca/imagegen-modelfile

parth/agent-skills

parth/agent-allowlist

parth/signed-in-offline

parth/agents

parth/fix-context-chopping

improve-cloud-flow

parth/add-models-websearch

parth/prompt-renderer-mcp

jmorganca/native-settings

jmorganca/download-stream-hash

jmorganca/client2-rebased

brucemacd/oai-chat-req-multipart

jessegross/multi_chunk_reserve

grace/additional-omit-empty

grace/mistral-3-large

mxyng/tokenizer2

mxyng/tokenizer

jessegross/flash

hoyyeva/windows-nacked-app

mxyng/cleanup-attention

grace/deepseek-parser

hoyyeva/remember-unsent-prompt

parth/add-lfs-pointer-error-conversion

parth/olmo2-test2

hoyyeva/ollama-launchagent-plist

nicole/olmo-model

parth/olmo-test

mxyng/remove-embedded

parth/render-template

jmorganca/intellect-3

parth/remove-prealloc-linter

jmorganca/cmd-eval

nicole/nomic-embed-text-fix

mxyng/lint-2

hoyyeva/add-gemini-3-pro-preview

hoyyeva/load-model-list

mxyng/expand-path

mxyng/environ-2

hoyyeva/deeplink-json-encoding

parth/improve-tool-calling-tests

hoyyeva/conversation

hoyyeva/assistant-edit-response

hoyyeva/thinking

origin/brucemacd/invalid-char-i-err

parth/improve-tool-calling

jmorganca/required-omitempty

grace/qwen3-vl-tests

mxyng/iter-client

parth/docs-readme

nicole/embed-test

pdevine/integration-benchstat

parth/remove-generate-cmd

parth/add-toolcall-id

mxyng/server-tests

jmorganca/glm-4.6

jmorganca/gin-h-compat

drifkin/stable-tool-args

pdevine/qwen3-more-thinking

parth/add-websearch-client

nicole/websearch_local

jmorganca/qwen3-coder-updates

grace/deepseek-v3-migration-tests

mxyng/fix-create

jmorganca/cloud-errors

pdevine/parser-tidy

revert-12233-parth/simplify-entrypoints-runner

parth/enable-so-gpt-oss

brucemacd/qwen3vl

jmorganca/readme-simplify

parth/gpt-oss-structured-outputs

revert-12039-jmorganca/tools-braces

mxyng/embeddings

mxyng/gguf

mxyng/benchmark

mxyng/types-null

parth/move-parsing

mxyng/gemma2

jmorganca/docs

mxyng/16-bit

mxyng/create-stdin

pdevine/authorizedkeys

mxyng/quant

parth/opt-in-error-context-window

brucemacd/cache-models

brucemacd/runner-completion

jmorganca/llama-update-6

brucemacd/benchmark-list

brucemacd/partial-read-caps

parth/deepseek-r1-tools

mxyng/omit-array

parth/tool-prefix-temp

brucemacd/runner-test

jmorganca/qwen25vl

brucemacd/model-forward-test-ext

parth/python-function-parsing

jmorganca/cuda-compression-none

drifkin/num-parallel

drifkin/chat-truncation-fix

jmorganca/sync

parth/python-tools-calling

drifkin/array-head-count

brucemacd/create-no-loop

parth/server-enable-content-stream-with-tools

qwen25omni

mxyng/v3

brucemacd/ropeconfig

jmorganca/silence-tokenizer

parth/sample-so-test

parth/sampling-structured-outputs

brucemacd/doc-go-engine

parth/constrained-sampling-json

jmorganca/mistral-wip

brucemacd/mistral-small-convert

parth/sample-unmarshal-json-for-params

brucemacd/jomorganca/mistral

pdevine/bfloat16

jmorganca/mistral

brucemacd/mistral

pdevine/logging

parth/sample-correctness-fix

parth/sample-fix-sorting

jmorgan/sample-fix-sorting-extras

jmorganca/temp-0-images

brucemacd/parallel-embed-models

brucemacd/shim-grammar

jmorganca/fix-gguf-error

bmizerany/nameswork

jmorganca/faster-releases

bmizerany/validatenames

brucemacd/err-no-vocab

brucemacd/rope-config

brucemacd/err-hint

brucemacd/qwen2_5

brucemacd/logprobs

brucemacd/new_runner_graph_bench

progress-flicker

brucemacd/forward-test

brucemacd/go_qwen2

pdevine/gemma2

jmorganca/add-missing-symlink-eval

mxyng/next-debug

parth/set-context-size-openai

brucemacd/next-bpe-bench

brucemacd/next-bpe-test

brucemacd/new_runner_e2e

brucemacd/new_runner_qwen2

pdevine/convert-cohere2

brucemacd/convert-cli

parth/log-probs

mxyng/next-mlx

mxyng/cmd-history

parth/templating

parth/tokenize-detokenize

brucemacd/check-key-register

bmizerany/grammar

jmorganca/vendor-081b29bd

mxyng/func-checks

jmorganca/fix-null-format

parth/fix-default-to-warn-json

jmorganca/qwen2vl

jmorganca/no-concat

parth/cmd-cleanup-SO

brucemacd/check-key-register-structured-err

parth/openai-stream-usage

parth/fix-referencing-so

stream-tools-stop

jmorganca/degin-1

brucemacd/install-path-clean

brucemacd/push-name-validation

brucemacd/browser-key-register

jmorganca/openai-fix-first-message

jmorganca/fix-proxy

jessegross/sample

parth/disallow-streaming-tools

dhiltgen/remove_submodule

jmorganca/ga

jmorganca/mllama

pdevine/newlines

pdevine/geems-2b

jmorganca/llama-bump

mxyng/modelname-7

mxyng/gin-slog

mxyng/modelname-6

jyan/convert-prog

jyan/quant5

paligemma-support

pdevine/import-docs

jmorganca/openai-context

jyan/paligemma

jyan/p2

jyan/palitest

bmizerany/embedspeedup

jmorganca/llama-vit

brucemacd/allow-ollama

royh/ep-methods

royh/whisper

mxyng/api-models

mxyng/fix-memory

jyan/q4_4/8

jyan/ollama-v

royh/stream-tools

roy-embed-parallel

bmizerany/hrm

revert-5963-revert-5924-mxyng/llama3.1-rope

royh/embed-viz

jyan/local2

jyan/auth

jyan/local

jyan/parse-temp

jmorganca/template-mistral

jyan/reord-g

royh-openai-suffixdocs

royh-imgembed

royh-embed-parallel

jyan/quant4

royh-precision

jyan/progress

pdevine/fix-template

jyan/quant3

pdevine/ggla

mxyng/update-registry-domain

jmorganca/ggml-static

mxyng/create-context

jyan/v0.146

mxyng/layers-from-files

build_dist

bmizerany/noseek

royh-ls

royh-name

timeout

mxyng/server-timestamp

bmizerany/nosillyggufslurps

royh-params

jmorganca/llama-cpp-7c26775

royh-openai-delete

royh-show-rigid

jmorganca/enable-fa

jmorganca/no-error-template

jyan/format

royh-testdelete

bmizerany/fastverify

language_support

pdevine/ps-glitches

brucemacd/tokenize

bruce/iq-quants

bmizerany/filepathwithcoloninhost

mxyng/split-bin

bmizerany/client-registry

jmorganca/if-none-match

native

jmorganca/native

jmorganca/batch-embeddings

jmorganca/initcmake

jmorganca/mm

pdevine/showggmlinfo

modenameenforcealphanum

bmizerany/modenameenforcealphanum

jmorganca/done-reason

jmorganca/llama-cpp-8960fe8

ollama.com

bmizerany/filepathnobuild

bmizerany/types/model/defaultfix

rmdisplaylong

nogogen

bmizerany/x

modelfile-readme

bmizerany/replacecolon

jmorganca/limit

jmorganca/execstack

jmorganca/replace-assets

mxyng/tune-concurrency

jmorganca/testing

whitespace-detection

jmorganca/options

upgrade-all

scratch

cuda-search

mattw/airenamer

mattw/allmodelsonhuggingface

mattw/quantcontext

mattw/whatneedstorun

brucemacd/llama-mem-calc

mattw/faq-context

mattw/communitylinks

mattw/noprune

mattw/python-functioncalling

rename

mxyng/install

pulse

remove-first

editor

mattw/selfqueryingretrieval

cgo

mattw/howtoquant

api

matt/streamingapi

format-config

mxyng/extra-args

shell

update-nous-hermes

cp-model

upload-progress

fix-unknown-model

fix-model-names

delete-fix

insecure-registry

ls

deletemodels

progressbar

readme-updates

license-layers

skip-list

list-models

modelpath

matt/examplemodelfiles

distribution

go-opts

1 Participants

Notifications

Due Date

No due date set.

Dependencies

No dependencies set.

Reference: github-starred/ollama#68640