-
**ISSUE TRANSFER: Optimum repository -> https://github.com/huggingface/optimum/issues/555**
This issue is about the working group specially created for this task. If you are interested in helping o…
-
### System Info
- `transformers` version: 4.30.2
- Platform: macOS-13.4.1-arm64-arm-64bit
- Python version: 3.11.4
- Huggingface_hub version: 0.16.3
- Safetensors version: 0.3.1
- PyTorch versio…
-
Implement a training function to fine-tune pre-trained Helsinki models. This function should accept raw files and their corresponding translated files from the ```load_data``` function.
Implement a…
yaeso updated
7 months ago
-
Hi,
Do you reckon you could guide me on how to support MarianMT models ?
They can be exported to ONNX using : https://github.com/kcosta42/Marian-ONNX-Converter
There are ONNX models already avail…
-
Hi, in the paper it is mentioned that you use only russian as an intermediate language for backtranslation. But the code highlights the presence of two different augmentation strategies - en-de-en for…
ghost updated
7 months ago
-
In file _prompt_translator.py_
def MarianMT_trans(text, **lang_pair**):
replace to
def MarianMT_trans(text, **tar_lang**):
-
# Back Translation for Text Augmentation with Google Sheets
Learn how to augment existing labeled text data for free using Google Sheets.
[https://amitness.com/2020/02/back-translation-in-google-she…
-
I am using official example (https://huggingface.co/blog/dpo-trl) to optimize T5 using DP0. Dataset format is the same but I still get error: 'tuple' object has no attribute 'logits'
here is the …
Oxi84 updated
9 months ago
-
## 🐛 Bug:
When compiling the model with mlc_chat compile getting the error r_loops assertion error.
This is happening at op.gelu(self.fc1(hidden_states)) which i could know as
gelu: R.Tens…
-
**MarianMT architecture**
I found an interesting detail about MarianMT implementation in huggingface. There is no "Softmax" layer after "Linear" at the end of the model, despite the default architect…