- b0b66a7 Kelvin Resize nearest neighbors for int8 by Naveen Dodda · 8 months ago
- 321c904 Rewrite ConvPerChannelD1OD24_5x5 by David Gao · 8 months ago
- 0758fb8 Improve Conv2dD32 by Alex Van Damme · 8 months ago
- 008f0ae Restrict ConvS16B64K1xn to height 1 by Alex Van Damme · 8 months ago
- f22bdde Fix in_channel calculation in ConvS8D4 by Alex Van Damme · 8 months ago
- 108cf32 Remove extra offset addition from ElementwiseAdd by Alex Van Damme · 8 months ago
- 5190b46 Merge "Add test_status calls to benchmark_smc" by Alex Van Damme · 8 months ago
- 35a67de Merge "Create templated version of Swizzle." by Alex Van Damme · 8 months ago
- 109726f Add test_status calls to benchmark_smc by Alex Van Damme · 8 months ago
- 5496087 Merge "Implement DepthwiseConvS85x5D32 w/ adwconv" by Alex Van Damme · 8 months ago
- 18694e4 Merge "Two outputs per iteration in ConvPerChannelD1OD24_5x5" by Alex Van Damme · 8 months ago
- 2cebe31 Implement DepthwiseConvS85x5D32 w/ adwconv by Alex Van Damme · 8 months ago
- f427e85 Create templated version of Swizzle. by Derek Chow · 8 months ago
- 41f1cec Reduce register pressure in ConvS8W8D4. by Derek Chow · 8 months ago
- f353919 Two outputs per iteration in ConvPerChannelD1OD24_5x5 by Alex Van Damme · 8 months ago
- 9e0b2ba Merge "Reduce vld calls in ConvS8D4." by Derek Chow · 8 months ago
- 2f60851 Reduce vld calls in ConvS8D4. by Derek Chow · 8 months ago
- 87cc267 Merge "Process 16 channels per iteration in 1x1 d16" by Alex Van Damme · 8 months ago
- cf069b3 Process 16 channels per iteration in 1x1 d16 by Alex Van Damme · 8 months ago
- b5caeab Merge "Improve ElementwiseAddS8" by Alex Van Damme · 8 months ago
- 74f2f74 Create and apply interleaved output pipeline macros by Alex Van Damme · 8 months ago
- 8da6e53 5x5 DepthwiseConv, 4 outputs per loop iteration by Alex Van Damme · 9 months ago
- ac91807 Improve ElementwiseAddS8 by Alex Van Damme · 9 months ago
- d22d7d6 DepthwiseConv 3x3, 4 per iteration by Alex Van Damme · 9 months ago
- 13fe02a Create RUN_KERNEL macro for DepthwiseConvS8 by Alex Van Damme · 9 months ago
- 088841b Specialize 3x3, stride 1 DepthwiseConv by Alex Van Damme · 9 months ago
- 5106685 Vectorized Logistic for int8 by Alex Van Damme · 9 months ago
- b309494 Merge changes I1f0ec9d3,I55e0cf22 by Cindy Liu · 9 months ago
- db10d3b Merge changes from topic "nexus-benchmark" by Cindy Liu · 9 months ago
- 627a3ba Lint cleanup benchmark_kelvin.cc by Cindy Liu · 9 months ago
- 2ec00d9 Fix carry over benchmark mismatch count by Cindy Liu · 9 months ago
- 8ef6f11 Use 1x1 d=32n kernel more aggressively. by Derek Chow · 9 months ago
- aaf9d30 Update benchmark device rule to support tags by Cindy Liu · 9 months ago
- f6ec59d Format benchmarks.bzl by Cindy Liu · 9 months ago
- 1ad0284 Point wise convolution input_depth = 16 by Derek Chow · 9 months ago
- 5160bc6 Add specialized pointwise convolution. by Derek Chow · 9 months ago
- dd44caa Allow specifying benchmark arena size by Alex Van Damme · 9 months ago
- d2aca97 Add additional ops to benchmark by Alex Van Damme · 9 months ago
- 78538b3 Use dwconv on input_depth == 1, 5x5 by Alex Van Damme · 9 months ago
- 0d75a42 Refactor kernel selection. by Derek Chow · 9 months ago
- 2aeaaa8 Enable O2 optimizations. by Derek Chow · 9 months ago
- 5971487 Improvements to elementwise_add_{s8,s16} by Alex Van Damme · 9 months ago
- ca72b40 Improve ConvS8D4 by Alex Van Damme · 9 months ago
- cd3d0e3 3x3 DepthwiseConv w/ adwconv by Alex Van Damme · 9 months ago
- b1afda6 Specialize 5x5 DepthwiseConv by Alex Van Damme · 10 months ago
- a99c93d Add support for checking output in benchmark by Alex Van Damme · 10 months ago
- fd54c9e Relax output_depth for ConvS8D4 by Alex Van Damme · 10 months ago
- 86806b3 Add input_depth == 1 specialized convolution. by Derek Chow · 10 months ago
- 456ed80 Merge "Support running a different Kelvin benchmark binary" by Michael Brooks · 9 months ago
- 40a8300 Fix swizzling and rounding for depthwise conv by Alex Van Damme · 10 months ago
- e617f61 Relax restrictions on aconv-based Conv2D by Alex Van Damme · 10 months ago
- 3b8d3cb sw/kelvin: Refactoring TFLM Kelvin kernel optimization codes by Lun Dong · 10 months ago
- d7880e8 Add doc on tflm op support in Kelvin. by Derek Chow · 10 months ago
- 2d626f6 Support running a different Kelvin benchmark binary by Michael Brooks · 10 months ago
- b11ad17 Specialized int8 conv kernels by Alex Van Damme · 10 months ago
- be4ab97 WIP Depthwise conv for 2d kernels by Naveen Dodda · 10 months ago
- c62ec01 Add wall-time measurement to Benchmark by Alex Van Damme · 11 months ago
- c47ddc8 Increasing memory to support models for benchmarks. by Naveen Dodda · 11 months ago
- 1eb552e Updating resolver to support mobilenet models by Naveen Dodda · 11 months ago
- ff78446 Add support for using test data in model benchmarks by Michael Brooks · 11 months ago
- 5f67634 Merge "Add Support for Profiling TFLM Models in Benchmarks" by Michael Brooks · 11 months ago
- d884926 Fix hw test script to support extra flags by Cindy Liu · 11 months ago
- fe2b2a8 Add Support for Profiling TFLM Models in Benchmarks by Michael Brooks · 11 months ago
- a1e166f Pass through extra flags to the hardware sim by Julian Mullings-Black · 11 months ago
- d48c5a4 Patches for benchmark on ASIC by Alex Van Damme · 11 months ago
- 3bdf8f4 Add minstret to perf_counters by Alex Van Damme · 1 year, 1 month ago
- c0214c3 Add ecall/mret test by Alex Van Damme · 12 months ago
- 8d688e2 Fix Benchmark SMC Print Format by Michael Brooks · 11 months ago
- 03b3adf Additonal Benchmark Targets by Michael Brooks · 11 months ago
- dc03ded Add benchmark suite to sw/kelvin by Michael Brooks · 11 months ago
- 215d48a Enabling kelvin_test outside of sw/kelvin. by Naveen Dodda · 1 year ago
- 0b2b7fb Type-check around find_cc_toolchain result usage by Alex Van Damme · 1 year, 1 month ago
- 2b6d459 Mark `model_input` and `output_header` as NOLOAD by Alex Van Damme · 1 year, 3 months ago
- 0cc464f Make pooling_test "large" to adjust timeout by Julian Mullings-Black · 1 year, 1 month ago
- 24619c5 Add perf_counters test by Alex Van Damme · 1 year, 1 month ago
- 78e3bc3 Update .bazelversion to 6.2.1 by Alex Van Damme · 1 year, 2 months ago
- 3285c6a Changes to optimize max_pool for Kelvin int8 dtype by Naveen-Dodda · 1 year, 2 months ago
- 0c9118e Add riscv CSR tests, enable zicsr by Alex Van Damme · 1 year, 3 months ago
- 4054485 Merge "sw/kelvin: revamp layout to support setting model input data" by Sam Leffler · 1 year, 3 months ago
- c231ced sw/kelvin: revamp layout to support setting model input data by Sam Leffler · 1 year, 4 months ago
- 903c311 Consolidate kernel selection behind generic conv_per_channel_b64. by Derek Chow · 1 year, 3 months ago
- a5f129c Optimize 1x1 group conv. by Derek Chow · 1 year, 4 months ago
- 385ea39 Fix PrintfTraits print out by Cindy Liu · 1 year, 4 months ago
- 2095c66 Update vsran unsigned tests by Cindy Liu · 1 year, 4 months ago
- 5a2cf62 Add vst unit test by Cindy Liu · 1 year, 4 months ago
- 9c9167a Cleanup kelvin_isa test format warnings by Cindy Liu · 1 year, 4 months ago
- ac27a72 Add vstq unit test by Cindy Liu · 1 year, 4 months ago
- 76d4adc Add vsrn and vzip unit tests by Cindy Liu · 1 year, 4 months ago
- 503e5ce Add more vector unit tests by Cindy Liu · 1 year, 4 months ago
- 2fe3247 Add acset actr unit tests by Cindy Liu · 1 year, 4 months ago
- 6448eab Add unit tests by Cindy Liu · 1 year, 4 months ago
- 3ffc473 OSS cleanup on the host_tool by Cindy Liu · 1 year, 4 months ago
- 66f7e3c Fix contributing doc code review description by Cindy Liu · 1 year, 4 months ago
- 43879e4 OSS cleanup by Cindy Liu · 1 year, 4 months ago
- c76bc0c Add depthwise conv test. by Derek Chow · 1 year, 4 months ago
- 766e5af Add dilation support for depthwise convolution. by Derek Chow · 1 year, 4 months ago
- 16ffb60 8x16 Depthwise convolution for Kelvin. by Derek Chow · 1 year, 4 months ago
- f39cadd Helper methods for Conv2D kernel by Alex Van Damme · 1 year, 6 months ago
- fc5baaa Move soundstream TFLM example to the model repository by Cindy Liu · 1 year, 5 months ago
- e525be0 Remove remaining C++ from kelvin.h, include in hello_world by Alex Van Damme · 1 year, 5 months ago