-
Hello,
many thanks for this very nice piece of work!
I couldn't get the finetune/finetune_lora scripts to run on a freshly launched ubuntu ec2 instance without a substantial refactoring of the …
-
Hi!
I have a basic question, why the whole fine rune does need the eval dataset?
-
Hi when trying to run Phi-3-vision, I got this error. Has anyone seen this before?
Loading model...
Traceback (most recent call last):
File "D:\phi3-vision\Phi-3-vision-128k-instruct-onnx-cuda\…
-
I set the following parameters and executed finetune.sh
```
--bf16 False \
--tf32 False \
--disable_flash_attn2 True
```
But I encountered the following error when saving checkpoint
```
…
-
**Describe the bug**
What the bug is, and how to reproduce, better with screenshots(描述bug以及复现过程，最好有截图)
```
swift infer --model_type internvl2-8b-awq --infer_backend lmdeploy
```
```
WARNING:ro…
-
1、下载了Phi-3-Vision数据
2、设置example_image.sh中的参数PER_DEVICE_BATCH_SIZE=1，NUM_EPOCHS=5，训练集选择了celeba_image_train.json中的87条，测试集选择celeba_image_eval.json，训练的时候到87步出错：
20%|█████████████████████████████▊ …
-
It seems to be a bit slower than using the original phi3-vision model.
I even tried ISQ and its still slower.
The original takes about 5 to 6 seconds and, when using mistral.rs it takes about 9 to…
-
"one-piece swimsuit" seems to trigger a KeyError.
On a side note, it would be nice to have WD auto tagging as optional as it doesn't suit every dataset.
[2024-06-12 15:16:28,877] ERROR in app:…
-
当我phi3_vision训练集中含有仅有文本的样本时出现报错如下：
KeyError: 'pixel_values'
项目中对应代码如下：
if batch[0].get('pixel_values') is not None:
res['pixel_values'] = torch.concat([b['pixel_values'] for b in batch])
…
-
**Describe the bug**
deepspeed-zero3，lora_target_modules ALL，model_type phi3-vision-128k-instruct，多机多卡，在resume from checkpoint的时候，模型似乎无法加载。需要注意的是，此时的chekpoint文件夹内只包括lora相关的参数，但是报错显示模型在加载更多参数。
> Fi…