Fine tuning gpt2 | Transformers huggingface | conversational chatbot

Fine tuning gpt2 | Transformers huggingface | conversational chatbot | GPT2LMHeadModel

Рет қаралды 30,335

Күн бұрын

Hi welcome to new video where we will code in python for fine tuning gpt2 model provided by huggingface.
We will be using convai dataset for finetuning as well as GPT2Tokenizer for tokenising the inputs.
Code : github.com/Pawandeep-prog/fin...
Thanks for watching the video :)
timestamps:
0:00 - Getting dataset and directory ready
1:16 - Using pretrained GPT2
4:01 - preparing dataset using Class
14:00 - model training setup
20:00 - Training on colab and Resolving errors
29:00 - Making predictions
contact me :
facebook : / proogramminghub
twitter : / programming_hut
github : github.com/Pawandeep-prog
discord : / discord
linkedin : / programminghut
youtube : / programminghutofficial
chatgpt,finetune,gpt2,model,transformers,pytorch,torch,huggingface,convai,dataset,dataloader,finetuning,deep learning,machine learning

Пікірлер: 64

@bennievaneeden2720 Жыл бұрын

Great video. Glad to see people actually debugging their code. It really helps to better grasp whats going on

@plashless3406 11 ай бұрын

This is amazing. Really appreciate your efforts, brother.

@ilyas8523 Жыл бұрын

great video, I was stuck on some steps until I found your video.

@kaenovama Жыл бұрын

Ayy, this is what I've been looking for

@rishabhramola448 Жыл бұрын

Really helpful!!! Deserve more views

@fp-mirzariyasatali1985 Жыл бұрын

Its great to try multiple techniques

@ammar46 9 ай бұрын

Nothing is generated for me. I used exact code as you used. Can you help understand why this is happening??

@ucduyvo4552 Жыл бұрын

Great Video, thanks

@shubhamkumar8093 Жыл бұрын

Very Informative 😊

@gingerbread_GB Жыл бұрын

This is a very important learning point. When I was watching your video, I agreed with you that it was not good the bot kept saying "I'm a very experienced person." But later on, when I dove into your json file, I realized that the dialogue in the file was bad. When I searched for "experienced" in the json, I got 61 hits, with the same "I'm not sure. I'm a very experienced person." over and over again. So the model worked, your code worked, the bot learns exactly what you taught it in the dataset. If your dataset is bad, your bot will talk nonsense.

@programming_hut Жыл бұрын

I am planning to release series of about 30 days long all about embedding langchain datasets huggingface and these llms CurrentlyI am extensively researching about these and will surely upload. And thanks for your intuition I will look about this too.

@gingerbread_GB Жыл бұрын

@@programming_hut Thanks for the response, I'm an amateur learning about this stuff too. I wish you can do a tutorial on LoRA with Huggingface transformers. for us at home, we don't have access to powerful GPUs to train big models. LORA make it more feasible. Another big part I'm struggling with is how to get big data. There are plenty of datasets out there scraped from the internet, but they are mostly filled with trash. Like inappropriate languages, messy grammar. May that's something you can talk about.

@programming_hut Жыл бұрын

Sure I will try to cover after some research

@georgekokkinakis7288 11 ай бұрын

What if we don't add the < , and to the training data? I have a dataset where each sample is formated as [context] [user_question] [bot_answer] and each sample is separated from the next one by an empty line. I am using a pretrained model lighteternal/gpt2-finetuned-greek

@sharathmandadi Жыл бұрын

what changes required to use this code for squad Question Answering training

@user-fn4ri4gf8r Жыл бұрын

great video. just wondering why u used

@sanketgaikwad996 Жыл бұрын

thanks a lot bro

@rakesh2233 9 ай бұрын

Great video, thanks. I am a beginner studying about these LLM's. I have a small doubt, I have seen people use different data formats to fine-tune different LLMs. For example, the following format can be used for Llama-2: { "instruction": "", "input": "", "output": "" } and sometimes the format below is used for chatglm2-6b: { "content": "", "summary": "" } Is it related to what format was used for pre-training or actually both can be used for different llms, how do I organize my custom data if I want to fine-tune a llm?

@elibrignac8050 Жыл бұрын

Epic video, just one question: After saving the model, how can I load it and make inferences.

@lekhwar Жыл бұрын

Great

@amitvyas7905 Жыл бұрын

Hi! It was a good video. I would like to know once the model is trained how can we check the accuracy? Can you generate a ROUGE score? That way we know how goo or bad the model is.

@sanvisundarrajan 8 ай бұрын

Hey , I debugged the same exact way .. the bot does'nt reply though. Can someone help with this ?

@Websbird Ай бұрын

Great help - Can you recommend any editor for windows just like yours? and is that possible for you to create a colab or kaggle for same project?

@muntazkaleem7977 Жыл бұрын

Nice work

@programming_hut Жыл бұрын

Thanks

@arjunparasher7005 Жыл бұрын

🙌🙌

@manthanghonge007 3 ай бұрын

hey can you tell me how to connect this gpt 2 model from front end

@miningahsean886 Жыл бұрын

Good Video Sir, where did you source your dataset?

@kotcraftchannelukraine6118 Жыл бұрын

Is it possible to fine tune OPT-125M or GPT-Neo-125M using this?

@MrJ3 Жыл бұрын

It's using the HuggingFace API so it's easy to swap models as long as the models support training on the task you are interested in. Just swap the model name out.

@coldbubbyguy6206 Жыл бұрын

Is it normal for the training to get stuck at 0% if I only have access to CPU?

@insigniamalignia Жыл бұрын

can't you use google colab? it's free

@tobi418 Жыл бұрын

Hello, thank you for sharing your code. I have question, Why don't you use Trainer class of HuggingFaces Transformers library for training?

@programming_hut Жыл бұрын

Nvm I just wanted to try this way

@tobi418 Жыл бұрын

@@programming_hut is there any advantage this way? Can I use Trainer class? Also can I use TextDataset class for import and tokenize my dataset?

@jackeyhua 2 ай бұрын

Hi there, very good video, really appreciate. Currently we are facing a problem that the input token does not generate any bot output, like . Can you help figure it out?

@user-eg7hi9rl7i Жыл бұрын

I wonder where I can find more dataset for finetuning a chatbot GPT-2, if anybody have idea please tell me, thanks.

@programming_hut Жыл бұрын

You can use chatgpt itself to generate more data You can look at alpaca model”s dataset

@nicandropotena2682 Жыл бұрын

Is it possible to train this model with the history of the conversation too? to keep track of what user said, in order to mantain a logical sense to the conversation.

@programming_hut Жыл бұрын

Working on that would probably find out and then will make tutorial.

@nitron3006 Жыл бұрын

after i run your code training doesn't work, it just kept on 0% | | 0/12 [00:00

@joshwaphly 3 ай бұрын

same error try chaning GPU to T4 in Run time tab Run selection, thanks alot to @programming_hut. You have enlightened me

@quantumjun Жыл бұрын

if finetuning needs to add a new layer?

@programming_hut Жыл бұрын

Fine-tuning trains a pretrained model on a new dataset without training from scratch Now it’s your choice to add or remove layer

@quantumjun Жыл бұрын

@@programming_hut thank you very much

@quantumjun Жыл бұрын

I reckon if we add a new layer for chatbot, it may overfit the chat data?

@quantumjun Жыл бұрын

since we don’t have enough data for that?

@neel_aksh Жыл бұрын

Can i use this with gpt-2-simple

@ucduyvo4552 Жыл бұрын

Thanhs sir, but the language in the fine tuning gpt2 video is English, how about the languages different English

@programming_hut Жыл бұрын

I might work on that and will try making video for it…

@UncleDavid 11 ай бұрын

why u breathing so fast? are u nervous?

@programming_hut 11 ай бұрын

😂 you pointed / I had cold 🥶 at that time

@mohammedismail6872 Жыл бұрын

bro pleasse place your mic away from keyboard.

@robosergTV Жыл бұрын

please speak faster, video too slow.

@programming_hut 11 ай бұрын

KZfaq has feature to make it fast in playback speed 😬

@balarajagopalan4981 Жыл бұрын

Talk slowly, man. Can't understand what you're saying.

@programming_hut Жыл бұрын

Sorry but you can reduce the speed

@rigeshyogi Жыл бұрын

@@programming_hut If you want to teach something, speak slowly and clearly so that you're understood. Otherwise, you're not going to reach a broader audience.

@programming_hut Жыл бұрын

Sure will keep in mind next time

@sloowed_reveerb Жыл бұрын

I'm not an English speaker and I understood everything! (and I'm not from India 😅😀) You can turn on subtitles if you're struggling :) Amazing tutorial btw, thanks to the author!

@programming_hut Жыл бұрын

thanks so kind of you