waqasm86/llamatelemetry-binaries

llamatelemetry Binaries v0.1.0

Pre-compiled CUDA binaries for llamatelemetry - CUDA-first OpenTelemetry Python SDK for LLM inference observability.

📦 Available Binaries

Version	File	Size	Target Platform	SHA256
v0.1.0	llamatelemetry-v0.1.0-cuda12-kaggle-t4x2.tar.gz	1.4 GB	Kaggle 2× Tesla T4, CUDA 12.5	31889a86116818be5a42a7bd4a20fde14be25f27348cabf2644259625374b355

🚀 Auto-Download (Recommended)

These binaries are automatically downloaded when you install llamatelemetry:

# Install on Kaggle with GPU T4 × 2
pip install --no-cache-dir --force-reinstall \
    git+https://github.com/llamatelemetry/llamatelemetry.git@v0.1.0

On first import llamatelemetry, the package will:

1. Detect your GPU (Tesla T4 required)
2. Check for cached binaries in ~/.cache/llamatelemetry/
3. Download from HuggingFace CDN (this repo - fast, ~2-5 MB/s)
4. Fallback to GitHub Releases if needed
5. Verify SHA256 checksum: 31889a86116818be5a42a7bd4a20fde14be25f27348cabf2644259625374b355
6. Extract 13 binaries + libraries to package directory
7. Configure environment variables

📥 Manual Download

Using huggingface_hub

from huggingface_hub import hf_hub_download

binary_path = hf_hub_download(
    repo_id="waqasm86/llamatelemetry-binaries",
    filename="v0.1.0/llamatelemetry-v0.1.0-cuda12-kaggle-t4x2.tar.gz",
    cache_dir="/kaggle/working/cache"
)

print(f"Downloaded to: {binary_path}")

Direct Download URL

wget https://huggingface.co/waqasm86/llamatelemetry-binaries/resolve/main/v0.1.0/llamatelemetry-v0.1.0-cuda12-kaggle-t4x2.tar.gz

Verify Checksum

# Download checksum file
wget https://huggingface.co/waqasm86/llamatelemetry-binaries/resolve/main/v0.1.0/llamatelemetry-v0.1.0-cuda12-kaggle-t4x2.tar.gz.sha256

# Verify
sha256sum -c llamatelemetry-v0.1.0-cuda12-kaggle-t4x2.tar.gz.sha256

📊 Build Information

Property	Value
Version	0.1.0
CUDA Version	12.5
Compute Capability	SM 7.5 (Tesla T4)
llama.cpp Version	b7760 (commit 388ce82)
Build Date	2026-02-03
Target Platform	Kaggle dual Tesla T4 GPUs (2× 15GB VRAM)
Binaries Included	13 (llama-server, llama-cli, llama-bench, etc.)
Libraries	CUDA shared libraries + dependencies

🔧 What's Inside

The binary bundle contains:

Executables (13 binaries)

• llama-server - OpenAI-compatible API server
• llama-cli - CLI inference tool
• llama-bench - Benchmarking utility
• llama-quantize - Model quantization tool
• And 9 more utilities

Shared Libraries

• CUDA 12.5 shared libraries
• cuBLAS, cuDNN dependencies
• llama.cpp runtime libraries

🔗 Links

• GitHub Repository: https://github.com/llamatelemetry/llamatelemetry
• GitHub Releases: https://github.com/llamatelemetry/llamatelemetry/releases/tag/v0.1.0
• Installation Guide: KAGGLE_INSTALL_GUIDE.md
• Models Repository: https://huggingface.co/waqasm86/llamatelemetry-models
• Documentation: https://llamatelemetry.github.io (planned)

🎯 Supported Platforms

Platform	GPU	CUDA	Status
Kaggle Notebooks	2× Tesla T4 (SM 7.5)	12.5	✅ Supported
Google Colab	Tesla T4 (SM 7.5)	12.x	🔄 Planned (v0.2.0)
Local Workstation	Tesla T4, RTX 4000+	12.x+	🔄 Planned (v0.2.0)
Other GPUs	SM < 7.5	Any	❌ Not supported

📄 License

MIT License - See LICENSE

🆘 Troubleshooting

Binary Download Fails

1. Check internet connection in Kaggle notebook settings
2. Retry import: import llamatelemetry (automatic retry logic)
3. Manual download: Use hf_hub_download() method above
4. GitHub fallback: Binaries also available at GitHub Releases

GPU Not Detected

from llamatelemetry import check_cuda_available, get_cuda_device_info

print(f"CUDA Available: {check_cuda_available()}")
print(f"GPU Info: {get_cuda_device_info()}")

Expected on Kaggle T4 × 2:

CUDA Available: True
GPU Info: {'gpu_name': 'Tesla T4', 'cuda_version': '12.5', 'compute_capability': '7.5'}

Incompatible GPU Error

llamatelemetry v0.1.0 requires Tesla T4 (SM 7.5) or newer. If you see "GPU compute capability < 7.5", you're running on an incompatible GPU.

Solution: Use Kaggle with "GPU T4 × 2" accelerator setting.

---

Maintained by: waqasm86
Version: 0.1.0
Last Updated: 2026-02-03
Status: Active Development