Whisper for Bangla & regional dialects
বাংলা ও আঞ্চলিক ভাষার জন্য WhisperFine-tune Whisper on FLEURS Bengali, then measure word error per region (Sylheti, Chittagong, Noakhali…).
Open this notebook ↗Read the short write-up →Copied
AI computing for LLM training and fine-tuning
USE CASE
Large-language-model work is usually constrained by AI Computer memory, tensor throughput and training time. ComputeBD is measuring demand for accelerator classes that can support experiments from compact fine-tunes to memory-heavy frontier models.
COMMON WORKLOADS
Where AI Computer acceleration can help
- Fine-tuning open-source LLMs
- Transformer training
- LoRA / QLoRA experiments
- Embedding and retrieval pipelines
- Multi-experiment research workflows
START IN ONE CLICK
Ready-made notebooks for this
Shared by the ComputeBD team. Open one, change it to fit your work, then run it on an AI Computer — billed by the minute.
Bangla dialect classifier (BanglaBERT)
আঞ্চলিক বাংলা চেনার ক্লাসিফায়ার (BanglaBERT)Train a text classifier on the BD-Dialect dataset; compare standard Bangla BERT models and their licences.
Open this notebook ↗Read the short write-up →Copied
RECOMMENDED AI COMPUTER CLASSES
NVIDIA H200
৳1,390*/hrHopper141 GBSXM
Memory-heavy LLMs, retrieval, frontier research
NVIDIA H100
৳1,047*/hrHopper80 GBSXM
Large model training, transformer workloads, HPC
NVIDIA A100
৳638*/hrAmpere80 GBSXM
LLM fine-tuning, research, high-throughput training
Tell us your workload and preferred AI Computer.
Join waitlist ↗FREE TOOLS
Plan it before you start
Can I run this model?
Memory (VRAM) needed for Llama, Qwen, Mistral, Stable Diffusion, FLUX… — which AI Computer fits, and the cost per hour in taka.
→Fine-tuning cost estimatorModel size, data and epochs → hours and cost in ৳. Print a one-page estimate for your supervisor or lab.
→Compare AI ComputersPick two to four AI Computers side by side: memory, speed, the largest model each can run, and the cost in taka.
→