-
I want to use dpo ft qwen-chat-14b, but I meet the error. The input(q, k, y) type of flash-attention in qwen has to be set as float16 or bfloat16, but in dpo_trainer the type is float32. If I turn of…
-
###yaml
```
model:
name: bert
framework: onnxrt_integerops
device: cpu
quantization:
approach: post_training_dynamic_quant
tuning:
accuracy_criterion:
relative: 0.01
e…
-
### Search before asking
- [X] I have searched the YOLOv8 [issues](https://github.com/ultralytics/ultralytics/issues) and [discussions](https://github.com/ultralytics/ultralytics/discussions) and fou…
-
### Search before asking
- [X] I have searched the YOLOv5 [issues](https://github.com/ultralytics/yolov5/issues) and found no similar bug report.
### YOLOv5 Component
Export
### Bug
Hello mainta…
-
### Prerequisites
- [X] I have read the [documentation](https://hf.co/docs/autotrain).
- [X] I have checked other issues for similar problems.
### Backend
Hugging Face Space/Endpoints
### Interfac…
-
脚本:
CUDA_VISIBLE_DEVICES=6,7 \
torchrun \
--nproc_per_node=$nproc_per_node \
--master_port=9901 \
examples/pytorch/llm/llm_sft.py \
--model_id_or_path ZhipuAI/chatglm3-6b \
…
bgtii updated
4 months ago
-
最终报错:
[INFO|training_args.py:1345] 2023-11-12 17:25:17,617 >> Found safetensors installation, but --save_safetensors=False. Safetensors should be a preferred weights saving format due to security and…
-
I want to modify the code of internlm-llava-7b, such as:
1. Change the vision encoder from CLIP ViT to DINO v2. They share the same architecture (ViT) but have different weights.
2. Modify the cod…
-
Hi, I am new to `peft`. I am wondering whether the `peft` supports use singular value decomposition to replace the orignal linear module. I find this is similar to the funciton of lora, but different …
-
### Reminder
- [X] I have read the README and searched the existing issues.
### Reproduction
Issue #2420 did not fix as of c901aa6.
Full error output
```cmd
[2024-03-12 08:50:54,244] torch.d…