Published in ACL Findings, 2024
We present PIXAR, the first pixel-based autoregressive LLM that understands and generates text-in-images, reaching GPT-2–level performance without relying on symbolic tokenization.
Published in SPIE, 2021
We present an attention-enhanced U-Net model integrating Squeeze-and-Excitation and CBAM modules to improve MRI brain tumor segmentation, delivered with an app featuring text-to-speech and chatbot support.
Published in ACL Findings, 2024
We present PIXAR, the first pixel-based autoregressive LLM that understands and generates text-in-images, reaching GPT-2–level performance without relying on symbolic tokenization.
Published in ACL, 2023
We propose simulating natural language feedback for interactive semantic parsing, enabling scalable training without costly human annotations and improving text-to-SQL error correction.
Published in ACL Findings, 2024
We present PIXAR, the first pixel-based autoregressive LLM that understands and generates text-in-images, reaching GPT-2–level performance without relying on symbolic tokenization.
slurm_gpu_report is a Bash command that summarizes GPU allocation and job usage in a Slurm cluster. It provides three views (nodes, jobs, users) and can export results to CSV.
Published in SPIE, 2021
We present an attention-enhanced U-Net model integrating Squeeze-and-Excitation and CBAM modules to improve MRI brain tumor segmentation, delivered with an app featuring text-to-speech and chatbot support.
HyperArgs is a typed configuration library for Python. It lets you define hyperparameters and settings as strongly typed Python classes, parse them from multiple sources (command line, JSON, TOML, YAML), enforce dependency constraints, and monitor field changes with decorators.
Published:
Tags: news
I’m very excited to be back in Edinburgh and to see the beautiful Edinburgh Futures Institute building. It feels inspiring to return to the city and begin a new chapter of research and discovery here.
HyperArgs is a typed configuration library for Python. It lets you define hyperparameters and settings as strongly typed Python classes, parse them from multiple sources (command line, JSON, TOML, YAML), enforce dependency constraints, and monitor field changes with decorators.
Published in ACL, 2023
We propose simulating natural language feedback for interactive semantic parsing, enabling scalable training without costly human annotations and improving text-to-SQL error correction.
slurm_gpu_report is a Bash command that summarizes GPU allocation and job usage in a Slurm cluster. It provides three views (nodes, jobs, users) and can export results to CSV.
Published in ACL Findings, 2024
We present PIXAR, the first pixel-based autoregressive LLM that understands and generates text-in-images, reaching GPT-2–level performance without relying on symbolic tokenization.
Published in ACL Findings, 2024
We present PIXAR, the first pixel-based autoregressive LLM that understands and generates text-in-images, reaching GPT-2–level performance without relying on symbolic tokenization.
HyperArgs is a typed configuration library for Python. It lets you define hyperparameters and settings as strongly typed Python classes, parse them from multiple sources (command line, JSON, TOML, YAML), enforce dependency constraints, and monitor field changes with decorators.
slurm_gpu_report is a Bash command that summarizes GPU allocation and job usage in a Slurm cluster. It provides three views (nodes, jobs, users) and can export results to CSV.