Comments (2)
Hi @Zcchill thanks for raising the question. This is expected - we just tokenize them as normal text. You can add them as special tokens after initializing the tokenizer, but the embeddings of the new tokens will be initialized to be something not that meaningful (could be random or an average of some existing embeddings) and might hurt the performance.
from open-instruct.
Thanks a lot for getting back to me!
from open-instruct.
Related Issues (20)
- Add support for resuming from checkpoint HOT 1
- llama reproduce issue HOT 10
- ignore question tokens in finetune.py HOT 1
- About the training seed HOT 2
- Add Decontamination Script
- size mismatch error when running merge_lora.py HOT 1
- Add system if to chat template
- Does the sharegpt dataset script itself very slow?
- oobabooga / llama-tokenizer takes forever to read HOT 4
- Is the split_all function in split_sharegpt_conversations.py very slow? HOT 5
- why delete finetune_trainer.py? HOT 1
- Any strategies for writing function convert_(dataset_name)_data for a custom dataset on hugging face? HOT 3
- Error when run ./scripts/finetune_with_accelerate.sh HOT 6
- Precompute DPO logprobs HOT 2
- Could not build wheels for flash-attn, vllm, xformers
- Save message IDs and source datasets during data mixing
- Commands for model evaluation HOT 1
- missing sharegpt_html_cleaned_and_split_2048.json HOT 1
- Make open-instruct pip installable HOT 1
- Make faster docker build
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from open-instruct.