RichardErkhov/ethzanalytics_-_distilgpt2-tiny-conversational-gguf

Quantization made by Richard Erkhov.

Github

Discord

Request more models

distilgpt2-tiny-conversational - GGUF

• —Model creator: https://huggingface.co/ethzanalytics/
• —Original model: https://huggingface.co/ethzanalytics/distilgpt2-tiny-conversational/

Original model description: --- license: apache-2.0 tags:

• —text-generation
• —chatbot
• —dialogue
• —distilgpt2
• —gpt2
• —ai-msgbot

widget:

• —text: "I know you're tired, but can we go for another walk this evening?\nperson beta:\n\n" example_title: "walk"
• —text: "Have you done anything exciting lately?\nperson beta:\n\n" example_title: "activities"
• —text: "hey - do you have a favorite grocery store around here?\nperson beta:\n\n" example_title: "grocery"
• —text: "Can you take me for dinner somewhere nice this time?\nperson beta:\n\n" example_title: "dinner"
• —text: "What's your favorite form of social media?\nperson beta:\n\n" example_title: "social media"
• —text: "Hi, how are you?\nperson beta:\n\n" example_title: "greeting"
• —text: "I am the best; my sister is the worst. What am I?\nperson beta:\n\n" example_title: "sister"
• —text: "What do you call an alligator who's just had surgery to remove his left arm?\nperson beta:\n\n" example_title: "alligator"
• —text: "A man walks into a bar and asks for a drink. The bartender asks for $10, and he pays him $1. What did he pay him with?\nperson beta:\n\n" example_title: "dollar"
• —text: "What did I say was in the mailbox when it was actually in the cabinet?\nperson beta:\n\n" example_title: "mailbox"
• —text: "My friend says that she knows every language, but she doesn't speak any of them.. what's wrong with her?\nperson beta:\n\n" example_title: "language"

inference: parameters: minlength: 2 maxlength: 64 lengthpenalty: 0.7 norepeatngramsize: 2 dosample: True topp: 0.95 topk: 20 temperature: 0.3 repetitionpenalty: 3.5

distilgpt2-tiny-conversational

This model is a fine-tuned version of distilgpt2 on a parsed version of Wizard of Wikipedia. Persona alpha/beta framework designed for use with ai-msgbot. It achieves the following results on the evaluation set:

• —Loss: 2.2461

Model description

• —a basic dialogue model for conversation. It can be used as a chatbot.
• —check out a simple demo here

Intended uses & limitations

• —usage is designed for integrating with this repo: ai-msgbot
• —the main specific information to know is that the model generates whole conversations between two entities, person alpha and person beta. These entity names are used functionally as custom <bos> tokens to extract when one response ends and another begins.

Training and evaluation data

• —wizard of Wikipedia parsed, from parlAI

Training procedure

• —deepspeed + huggingface trainer, an example notebook is in ai-msgbot

Training hyperparameters

The following hyperparameters were used during training:

• —learning_rate: 2e-05
• —trainbatchsize: 32
• —evalbatchsize: 32
• —seed: 42
• —distributed_type: multi-GPU
• —gradientaccumulationsteps: 4
• —totaltrainbatch_size: 128
• —optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• —lrschedulertype: cosine
• —lrschedulerwarmup_ratio: 0.05
• —num_epochs: 30

Training results

Framework versions

• —Transformers 4.16.1
• —Pytorch 1.10.0+cu111
• —Tokenizers 0.11.0