List of all items
Structs
- buffers::BufferArena
- buffers::BufferSizes
- buffers::decode_meta::DecodeMetaLayout
- cuda_backend::AtlasCudaBackend
- cuda_backend::tensormap::TensorMap
- fast_weights::FastSafetensorsLoader
- gpu::DevicePtr
- gpu::GraphHandle
- gpu::KernelHandle
- kernel_args::KernelLaunch
- kernel_audit::AuditRow
- kernel_audit::FailureSplit
- kv_cache::KvCacheConfig
- kv_cache::PagedKvCache
- kv_spill::KvSpillManager
- launch_trace::Entry
- op_cache::OpCache
- prefix_cache::EvictedBlocks
- prefix_cache::InsertAcquired
- prefix_cache::NoPrefixCaching
- prefix_cache::PrefixMatch
- radix_tree::RadixTree
- run_metrics::RunMetrics
- sampler::Sampler
- sampler::SamplingParams
- weights::DeferredTensor
- weights::GgufLoader
- weights::SafetensorsLoader
- weights::WeightStore
- weights::WeightTensor
- weights::mlx_int8::MlxInt8Weight
- weights::mlx_int8::MlxQuantConfig
Enums
Traits
Functions
- buffers::q12_batched_scratch_bytes
- buffers::q12_batched_scratch_bytes_varlen
- buffers::q2_dequant_scratch_bytes
- cublaslt::bf16_gemm_act_weight_t
- cublaslt::fp8_gemm_act_weight_t_blkscaled
- cublaslt::fp8_gemm_act_weight_t_rowwise
- cublaslt::prewarm
- cuda_backend::cuda_free_memory_bytes
- cuda_backend::spawn_oom_watchdog
- cutlass::bf16_gemm_act_weight_t
- cutlass::nvfp4_gemm_bf16_act_weight_t
- cutlass::nvfp4_grouped_down
- cutlass::nvfp4_grouped_gate_up
- cutlass::nvfp4_grouped_gate_up_fused
- cutlass::pack_bf16_weight_to_nvfp4_t
- cutlass::pack_weight_sfb
- cutlass::transpose_nvfp4_packed_kton
- fast_weights::is_vision_tensor
- flashinfer::available
- flashinfer::ragged_prefill_bf16_hd128
- flashinfer::ragged_prefill_bf16_hd256
- gpu::baseline_free_bytes
- gpu::pack_kernel_args
- gpu::set_baseline_free_bytes
- kernel_args::div_ceil
- kernel_audit::audit_rows
- kernel_audit::classify_failures
- kernel_audit::record
- kernel_audit::render_kernel_table
- kernel_audit::seal
- kernel_audit::split_failures
- kernel_audit::unresolved_lookups
- kernel_audit::unresolved_report
- kernel_audit::unseal
- kv_dequant::dequant_4bit_block_to_bf16
- kv_dequant::dequant_fp8_to_bf16
- kv_dequant::dequant_turbo3_block_to_bf16
- kv_dequant::dequant_turbo8_block_to_bf16
- kv_dequant::e4m3_lut
- launch_trace::begin
- launch_trace::end_and_diff
- launch_trace::name_kernel
- launch_trace::on
- launch_trace::record
- pinned_hosts::is_pinned
- pinned_hosts::live_count
- pinned_hosts::register
- pinned_hosts::unregister
- prefix_cache::cache_hit_count
- prefix_cache::cache_hit_tokens_total
- prefix_cache::cache_miss_count
- prefix_cache::record_cache_hit
- prefix_cache::record_cache_miss
- progress::layer
- progress::phase
- progress::preflight
- progress::ready
- progress::shard_done
- progress::shard_start
- run_metrics::cache_counts_this_run
- run_metrics::metrics
- run_metrics::reset_for_new_run
- sampler::apply_dry_penalty
- sampler::apply_lz_penalty
- sampler::apply_penalties_and_bias
- sampler::argmax_bf16
- sampler::argmax_f32
- sampler::argmax_first_wins_f32
- sampler::last_entropy
- sampler::low_entropy_token_count
- sampler::sample_with_params
- sampler::sample_with_params_history
- sampler::sample_with_params_seeded
- sampler::total_sampled_token_count
- set_ssm_tail_midchunk
- ssm_tail_boundary
- ssm_tail_ckpt_enabled
- ssm_tail_midchunk_enabled
- weights::adapter::load_adapter_safetensors
- weights::auto_detect_weight_prefix
- weights::config_from_gguf_dir
- weights::find_gguf
- weights::is_ngram_table
- weights::mlx_int8::gemv_gate_up
- weights::parse_expert_index