-
我参考文档docs/Yuan2_fastchat.md,进行训练,我的数据格式如下:
[{"id": "identity_0",
"conversations": [
{"from": "human", "value": "嘿,Bob,你最近看过什么好看的电影吗?"},
{"from": "gpt", "value": "嗯,我上周刚看了《流浪地球2》,简直…
-
Goal is to create_dataset.py, where we take the dataframe and the images from scrap_notes.py and create a Dataset consisting of:
[
{
"id": "000000033471",
"image": "000000033471.jpg",
…
-
As of now, the code is built in a way that it stores the conversation into a DB on the server. This is fine, but useless at this stage. The first thing to do would be to allow the possibility to retri…
-
So this is a strange one. I am stumped.
In way, this is sort of like #416, but I confirmed that if Batch==1, then the problem does not occur. (See below)
My inference loop looks like this
```
…
-
Can I use a multi turn conversations about an image to finetune the model? as human gpt human gpt for instance? Or they must be divided into 2 conversations?
for example:
"id": "FOLDER/IM…
-
### Question
Hi,
Thank you for the comprehensive guidelines on fine-tuning LLaVA on custom datasets! The provided example for generating tag-style captions for Stable Diffusion has been very helpf…
-
I need a dashboard where I can analyze bot conversations with users. It also provides the ability to connect the bot at any time with a live human agent.
-
### Question
Hi, When I try to look into the json file for finetuning, I find things like:
```
{
"id": "VG_100K_2/2412354",
"image": "vg/VG_100K_2/2412354.jpg",
"convers…
-
### Extension
https://www.raycast.com/abielzulio/chatgpt
### Description
I often have questions about technical and mathematical things. The response then contains text similar to what can be seen …
-
I'm getting an error while using the conversational colab:
`: 'list' object has no attribute 'keys'`
I just replaced: `load_dataset("json", data_files = {"train" : "my_data.json"})`
To use a fi…