# ValueError: Unrecognized model identifier

**URL:** <https://discuss.huggingface.co/t/valueerror-unrecognized-model-identifier/50112>\
**Category:** Models\
**Created:** [August 9, 2023, 10:18am UTC](https://discuss.huggingface.co/t/valueerror-unrecognized-model-identifier/50112 "2023-08-09T10:18:48Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![Bass2015](https://avatars.discourse-cdn.com/v4/letter/b/bb73d2/32.png) [@Bass2015](https://discuss.huggingface.co/u/Bass2015)\
**Post date:** [August 9, 2023, 10:18am UTC](https://discuss.huggingface.co/t/valueerror-unrecognized-model-identifier/50112/1 "2023-08-09T10:18:48Z")

</div>

Hi there,  
Please, I need help! I really don’t understand why I’m getting this error. It happens when I try to load an AutoModel either with the model name or the model path. Any of this cases throws the error:

```auto
model = AutoModel.from_pretrained("thenlper/gte-large")
model = AutoModel.from_pretrained("intfloat/multilingual-e5-large")
model = AutoModel.from_pretrained('./local_directory_multilingual-e5-large/')

```

The most strange thing is that until one hour ago, everything was working fine. Yesterday I downloaded the multilingual e5 model and saved it locally. When trying to load it again, the error happens.

```auto
ValueError: Unrecognized model identifier in intfloat/multilingual-e5-large. Should contains one of 'bert', 'openai-gpt', 'gpt2', 'transfo-xl', 'xlnet', 'xlm', 'roberta', 'ctrl`

```

I’m completely lost with this. Any help?

---

<div class="post-metadata">

**Author:** ![Bass2015](https://avatars.discourse-cdn.com/v4/letter/b/bb73d2/32.png) [@Bass2015](https://discuss.huggingface.co/u/Bass2015)\
**Post date:** [August 10, 2023, 8:18am UTC](https://discuss.huggingface.co/t/valueerror-unrecognized-model-identifier/50112/2 "2023-08-10T08:18:32Z")

</div>

In case any has the same issue:

The problem apparently was that when I installed `transformers` with conda, the version 2.1.1 was installed. I removed the package and installed it with pip and no problem
