Qingyu Song

icon indicating a website link https://simmonssong.github.io/ icon indicating an email [email protected]

icon company The Chinese University of Hong Kong icon location Hong Kong icon location Ph.D. Candidate at CUHK

Results 3 comments of Qingyu Song

Question about the datasets used for training

@amiryanj Hi, Did you check the distributions of train-set, validation-set, and test-set? Are the proportions between different kinds of trajectories the same? For instance, turning left, turning right, or going...

How to predict a specific length of tokens?

`max_tokens` cannot ensure an exact number of predicted tokens. Sometimes, a model predicts less than `max_tokens `.

Misc. bug: I'm seeing gibberish output

Similar problem of Qwen2.5-3B-Instruct with Q2_K on Win laptops: https://github.com/ggml-org/llama.cpp/discussions/12378