**Author:**  [@sophiaaez](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Fu%2Fsophiaaez)%5C
**Post date:** [June 16, 2022, 7:44am UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Fhow-to-create-a-config-json-after-saving-a-model%2F10459%2F12 "2022-06-16T07:44:36Z")
That only works for models that are transformer native and not nn.Module/pytorch native, sadly.
---
**Author:**  [@BramVanroy](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Fu%2FBramVanroy)%5C
**Post date:** [June 16, 2022, 9:03am UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Fhow-to-create-a-config-json-after-saving-a-model%2F10459%2F13 "2022-06-16T09:03:52Z")
What is your use-case that you are using Transformers but not Transformers models? If you want to use the HF Trainer alongside with your own PyTorch model, I recommended to subclass the relevant classes, similar to PretrainedModel
>
And to use your own PretrainedConfig alongside of it.
---
**Author:**  [@Hosna](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Fu%2FHosna)%5C
**Post date:** [July 27, 2022, 10:53am UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Fhow-to-create-a-config-json-after-saving-a-model%2F10459%2F14 "2022-07-27T10:53:21Z")
I have a similar issue where I have my model’s (nn.module) weights and I want to convert it to be huggingface compatible model so that I can use hugging face models (as `.generate`). From the discussions I can see that I either have to retrain again while changing (`nn.module` to `PreTrained`) or to define my config.json file. If I wrote my config.json file what should I do next to load my torch model as huggingface one?
---
**Author:**  [@tanyaroosta](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Fu%2Ftanyaroosta)%5C
**Post date:** [August 4, 2022, 3:25am UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Fhow-to-create-a-config-json-after-saving-a-model%2F10459%2F15 "2022-08-04T03:25:30Z")
I am not sure from the discussion above, what the solution is. Can someone post their working example please?
---
**Author:**  [@thepurpleowl](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Fu%2Fthepurpleowl)%5C
**Post date:** [September 8, 2022, 12:59pm UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Fhow-to-create-a-config-json-after-saving-a-model%2F10459%2F16 "2022-09-08T12:59:53Z")
I am not sure whether this functionality exists at this moment.
---
**Author:**  [@nairajay2k](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Fu%2Fnairajay2k)%5C
**Post date:** [November 14, 2022, 2:15pm UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Fhow-to-create-a-config-json-after-saving-a-model%2F10459%2F17 "2022-11-14T14:15:49Z")
Folks,
I am trying out the code at [GitHub - aws-samples/aws-inferentia-huggingface-workshop: CMP314 Optimizing NLP models with Amazon EC2 Inf1 instances in Amazon Sagemaker](https://github.com/aws-samples/aws-inferentia-huggingface-workshop) .
For inferentia instance, I get a config.json not found in the cloudwatch logs and inference fails. The config.json file is present in the traced model tar.gz file for inferentia.
Please help me resolve this.
Thanks
Ajay
P.S the log message — W-9002-model\_1-stdout MODEL\_LOG - OSError: file /home/model-server/tmp/models/cb9491669c1f44c1a0763e8a62d9368e/config.json not found
---
**Author:**  [@Kason123](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Fu%2FKason123)%5C
**Post date:** [August 30, 2023, 7:25pm UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Fhow-to-create-a-config-json-after-saving-a-model%2F10459%2F18 "2023-08-30T19:25:19Z")
Was anyone able to resolve this issue ,i.e., converting a custom nn.Module to a huggingface compatible version?
---
**Author:**  [@Kason123](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Fu%2FKason123)%5C
**Post date:** [August 30, 2023, 8:21pm UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Fhow-to-create-a-config-json-after-saving-a-model%2F10459%2F19 "2023-08-30T20:21:05Z")
Did you find a solution/workaround for this issue?
---
**Author:**  [@abdouaziiz](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Fu%2Fabdouaziiz)%5C
**Post date:** [August 31, 2023, 9:06pm UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Fhow-to-create-a-config-json-after-saving-a-model%2F10459%2F20 "2023-08-31T21:06:26Z")
@Hosna You can push the config via : customModel.pretrained\_model.config.push\_to\_hub(repo\_id)
---
**Author:**  [@aframson](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Fu%2Faframson)%5C
**Post date:** [September 15, 2023, 3:08pm UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Fhow-to-create-a-config-json-after-saving-a-model%2F10459%2F21 "2023-09-15T15:08:24Z")
SO this worked for me , i
# imported
`from transformers.modeling_utils import PreTrainedModel ,PretrainedConfig`
and then in my class
```auto
class TransformerLanguageModel(PreTrainedModel):
def __init__ (self, config):
super(TransformerLanguageModel, self). __init__ (config)
self.token_embedding_table = nn.Embedding(config.vocab_size, config.hidden_size)
self.position_embedding_table = nn.Embedding(config.block_size, config.hidden_size)
self.transformer = nn.Transformer(
d_model=config.hidden_size,
nhead=config.num_attention_heads,
num_encoder_layers=config.num_hidden_layers,
num_decoder_layers=config.num_hidden_layers,
dim_feedforward=4 * config.hidden_size,
dropout=config.hidden_dropout_prob,
activation='gelu'
)
self.ln1 = nn.LayerNorm(config.hidden_size)
self.ln2 = nn.LayerNorm(config.hidden_size)
self.lm_head = nn.Linear(config.hidden_size, config.vocab_size)
```
after that you have to create a config variable
# Create a configuration object
```auto
config = PretrainedConfig(
vocab_size=1000, # Specify your vocabulary size
hidden_size=n_embd, # Use your embedding dimension
num_attention_heads=n_head,
num_hidden_layers=n_layer,
hidden_dropout_prob=dropout,
block_size=block_size
)
```
```auto
model = TransformerLanguageModel(config)
model.to(device)
now you can save the model
model.save_pretrained('./path_to_model/')
```
[Next page](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Fhow-to-create-a-config-json-after-saving-a-model%2F10459.md%3Fpage%3D2)