# Why do I get a "UserWarning: No valid model found at models"?

**URL:** <https://forum.rasa.com/t/why-do-i-get-a-userwarning-no-valid-model-found-at-models/56242>\
**Category:** Rasa Open Source\
**Created:** [November 19, 2022, 11:58pm UTC](https://forum.rasa.com/t/why-do-i-get-a-userwarning-no-valid-model-found-at-models/56242 "2022-11-19T23:58:54Z")\
**Posts on this page:** 9\
**Page:** 1

<div class="post-metadata">

**Author:** ![MelT](https://dub1.discourse-cdn.com/flex013/user_avatar/forum.rasa.com/melt/32/21530_2.png) [@MelT](https://forum.rasa.com/u/MelT)\
**Post date:** [November 19, 2022, 11:58pm UTC](https://forum.rasa.com/t/why-do-i-get-a-userwarning-no-valid-model-found-at-models/56242/1 "2022-11-19T23:58:54Z")

</div>

Hi all,

I try to deploy an already trained model with Okteto. However, after it says "Loading model models/latestmodel.tar.gz I get the message

> UserWarning: No valid model found at models!

Is there something I am doing wrong? I successfully deployed the model locally with docker and it can also access the model but why not when deploying in the cloud?

This is my Dockerfile:

```auto
FROM python:3.7.7-stretch AS BASE

RUN apt-get update \
    && apt-get --assume-yes --no-install-recommends install \
        build-essential \
        curl \
        git \
        jq \
        libgomp1 \
        vim

WORKDIR /app

# upgrade pip version
RUN pip install --no-cache-dir --upgrade pip

RUN pip install rasa==3.1

ADD . .

```

While this is my Docker-compose file:

```auto
version: '3.4'
services:

    
  rasa-server:
    image: rasa/rasa:3.1.1-full
    working_dir: /app
    build:
      context: ./
      dockerfile: Dockerfile
    command: bash -c "rasa run --model latestmodel.tar.gz --enable-api --cors \"*\" --debug --port 5006"
    
    ports:
    - '5006:5006'

  rasa-actions-server:
    image: rasa/rasa:3.1.1-full
    working_dir: /app
    build:
      context: ./
      dockerfile: Dockerfile
    command: ["rasa", "run", "actions"]
    ports:
    - '5055:5055'

```

---

<div class="post-metadata">

**Author:** ![nik202](https://dub1.discourse-cdn.com/flex013/user_avatar/forum.rasa.com/nik202/32/16098_2.png) [@nik202](https://forum.rasa.com/u/nik202)\
**Post date:** [November 20, 2022, 11:36am UTC](https://forum.rasa.com/t/why-do-i-get-a-userwarning-no-valid-model-found-at-models/56242/2 "2022-11-20T11:36:21Z")

</div>

@MelT hope this will help you

[![](https://img.youtube.com/vi/hZqoERJ-23o/maxresdefault.jpg "Build a Rasa Chatbot with Okteto") ](https://www.youtube.com/watch?v=hZqoERJ-23o)

Good luck 🤞

---

<div class="post-metadata">

**Author:** ![MelT](https://dub1.discourse-cdn.com/flex013/user_avatar/forum.rasa.com/melt/32/21530_2.png) [@MelT](https://forum.rasa.com/u/MelT)\
**Post date:** [November 20, 2022, 11:43am UTC](https://forum.rasa.com/t/why-do-i-get-a-userwarning-no-valid-model-found-at-models/56242/3 "2022-11-20T11:43:33Z")

</div>

Hi @nik202 they train a model in Okteto but do not deploy an already trained model

---

<div class="post-metadata">

**Author:** ![nik202](https://dub1.discourse-cdn.com/flex013/user_avatar/forum.rasa.com/nik202/32/16098_2.png) [@nik202](https://forum.rasa.com/u/nik202)\
**Post date:** [November 20, 2022, 11:50am UTC](https://forum.rasa.com/t/why-do-i-get-a-userwarning-no-valid-model-found-at-models/56242/4 "2022-11-20T11:50:28Z")

</div>

Ok @MelT what is your use case why you only want to deploy on okteto ?

---

<div class="post-metadata">

**Author:** ![MelT](https://dub1.discourse-cdn.com/flex013/user_avatar/forum.rasa.com/melt/32/21530_2.png) [@MelT](https://forum.rasa.com/u/MelT)\
**Post date:** [November 20, 2022, 11:53am UTC](https://forum.rasa.com/t/why-do-i-get-a-userwarning-no-valid-model-found-at-models/56242/5 "2022-11-20T11:53:19Z")

</div>

I found it easier than deploying on GoogleCloud 🙂I think I could also load a modal from Google Cloud Storage but the linked tutorials here [Model Storage](https://rasa.com/docs/rasa/model-storage/) do not really explain where one has to set the environment variables. Do you know a tutorial on how to fetch a model from GCS?

---

<div class="post-metadata">

**Author:** ![nik202](https://dub1.discourse-cdn.com/flex013/user_avatar/forum.rasa.com/nik202/32/16098_2.png) [@nik202](https://forum.rasa.com/u/nik202)\
**Post date:** [November 20, 2022, 12:06pm UTC](https://forum.rasa.com/t/why-do-i-get-a-userwarning-no-valid-model-found-at-models/56242/6 "2022-11-20T12:06:46Z")

</div>

@MelT currently I can’t share but maybe later

Ok for your solution you can check the docker container mean inspect it and then update your already trained model. But keep in mind the code should not be updated or even rasa version.

---

<div class="post-metadata">

**Author:** ![MelT](https://dub1.discourse-cdn.com/flex013/user_avatar/forum.rasa.com/melt/32/21530_2.png) [@MelT](https://forum.rasa.com/u/MelT)\
**Post date:** [November 20, 2022, 12:19pm UTC](https://forum.rasa.com/t/why-do-i-get-a-userwarning-no-valid-model-found-at-models/56242/7 "2022-11-20T12:19:04Z")

</div>

Could explain a bit in more detail what you mean by inspecting? What should I look for?

---

<div class="post-metadata">

**Author:** ![nik202](https://dub1.discourse-cdn.com/flex013/user_avatar/forum.rasa.com/nik202/32/16098_2.png) [@nik202](https://forum.rasa.com/u/nik202)\
**Post date:** [November 20, 2022, 12:33pm UTC](https://forum.rasa.com/t/why-do-i-get-a-userwarning-no-valid-model-found-at-models/56242/8 "2022-11-20T12:33:04Z")

</div>

@MelT if you want to inspect your docker you need such command

docker exec -u root -t -i actions\_server /bin/bash

actions\_server is your container name for seeing the running image or container use this command

docker ps

---

<div class="post-metadata">

**Author:** ![MelT](https://dub1.discourse-cdn.com/flex013/user_avatar/forum.rasa.com/melt/32/21530_2.png) [@MelT](https://forum.rasa.com/u/MelT)\
**Post date:** [November 20, 2022, 12:34pm UTC](https://forum.rasa.com/t/why-do-i-get-a-userwarning-no-valid-model-found-at-models/56242/9 "2022-11-20T12:34:12Z")

</div>

Okay but I think the problem is maybe not in docker as it is working locally