minnow
Fast LLaDA2.2 inference server
File Explorer
- config.toml
- api.md
- cuda-architecture.json
- cuda-architecture.md
- decode-length-sweep.json
- decode-length-sweep.md
- decode-performance.json
- decode-performance.md
- implementation.md
- int8-measurements.json
- int8-optimizations.md
- loading.json
- loading.md
- memory.md
- model-format.md
- nvfp4-optimizations.json
- nvfp4-optimizations.md
- nvfp4-profile.json
- nvfp4-profile.md
- prefix-cache.json
- prefix-cache.md
- quantization-performance.json
- quantization-performance.md
- results.json
- results.md
- rtx3090.md
- spark-performance.json
- spark-performance.md
- throughput-comparison.json
- throughput-comparison.md
- validation.md
- bench_attention.rs
- bench_expert_formats.rs
- bench_load.rs
- bench_nvfp4.rs
- bench_quantized.rs
- check_prefix_numerics.rs
- compare_attention.rs
- profile_head.rs
- trace_prefill_boundaries.rs
- bench_cuda_arch.py
- bench_models.py
- check_batching.py
- check_compatibility.py
- check_prefix_cache.py
- check_server.py
- compare.py
- copy_checkpoint.py
- layerwise_reference.py
- memory_guard.py
- memory_pressure.py
- quantized_reference.py
- reference.py
- reference_generate.py
- test_memory_guard.py
- weight_io.py
- writer.rs
- flash.cu
- flash.rs
- norm.rs
- nvfp4.cu
- nvfp4.rs
- nvfp4_ampere.cu
- predict.rs
- qkv.cu
- qkv.rs
- quantized.cu
- quantized.rs
- quantized_int8.cu
- routed_experts.rs
- routing.cu
- routing.rs
- workspace.rs
- batching.rs
- cache_tests.rs
- nvfp4.rs
- batching.rs
- output.rs
- request.rs
- tests.rs
- webui.rs
- allocation.rs
- buffer.rs
- reader.rs
- config.rs
- container.rs
- cuda.rs
- decode.rs
- kernels.cu
- lib.rs
- main.rs
- model.rs
- prefix.rs
- quant.rs
- server.rs
- tokenizer.rs
- weights.rs
- config.json
- input.json
- metadata.json
- model.safetensors
- reference.safetensors
- README.md
- container.rs
- cpu_quantized.rs
- decode_comparison_cases.json
- decode_natural_cases.json
- decode_throughput_cases.json
- decoding.rs
- generation_cases.json
- prefill_benchmark_ids.json
- reference.rs
- block_info.h
- dropout.h
- flash.h
- flash_fwd_kernel.h
- kernel_traits.h
- LICENSE
- LICENSE-CANDLE-APACHE
- LICENSE-CUTLASS
- mask.h
- philox.cuh
- README.md
- rotary.h
- softmax.h
- utils.h
- .gitignore
- AGENTS.md
- build.rs
- Cargo.lock
- Cargo.toml
- LICENSE
- NOTICE
- README.md
# Use via CDN
jsDelivrjsDelivr serves any public GitHub repository as a CDN with zero setup. Pick a version and a file to get a ready-to-paste link and snippet.
Link
Example
// repository documentation
Was this content helpful?
(0 ratings)