openvela
/
apps_mlearning_tflite-micro
0
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
apps_mlearning_tflite-micro/tensorflow/lite/micro/kernels
History
Xie Lin 12699e7952 Tflm: Add per-channel quantization support to FULLY_CONNECTED operator 
This commit fixes the inference accuracy issue in tflite-micro
by adding per-channel quantization support to the FULLY_CONNECTED operator.

Root Cause:
- The fork only supported per-tensor quantization (single scale/zero_point
  for entire weight tensor)
- Upstream added per-channel quantization (independent scale/zero_point per
  output channel) which provides better accuracy

Changes:
1. tensorflow/lite/micro/kernels/fully_connected.h
   - Added is_per_channel flag
   - Added per_channel_output_multiplier array
   - Added per_channel_output_shift array

2. tensorflow/lite/micro/kernels/fully_connected_common.cc
   - Updated quantization parameter calculation to support per-channel

3. tensorflow/lite/micro/kernels/fully_connected.cc
   - Added conditional execution path for per-channel quantization
   - Preserved TFLITE_MODEL_COMPILER customization

Signed-off-by: Xie Lin <xielin3@xiaomi.com>
 		2026-04-20 14:48:10 +08:00
..
arc_mli Revert "Compute output shapes for some kernels (#2356)" (#2390) 2026-04-20 14:47:58 +08:00
beco tflite-micro:fix beco conv crash issue 2026-04-20 14:48:08 +08:00
ceva Add Q16x8 Depthwise Conv Support (#2140) 2023-08-01 23:40:18 +00:00
cmsis_nn tflite-micro:fix cmake cmsis nn add.cc compile error 2026-04-20 14:48:08 +08:00
delegate tflite: Delegate mechanism 2026-04-20 14:48:05 +08:00
ethos_u tflm: use system call for ethos-u driver 2026-04-20 14:48:10 +08:00
test_data_generation refactor: fix signed vs. unsigned comparison compiler warnings (#2064) 2023-06-16 18:29:27 +00:00
testdata Replace tensorflow-cpu usage with tensorflow (#2396) 2026-04-20 14:48:01 +08:00
xtensa tflite-micro: add xtensa strided_slice operator compile function 2026-04-20 14:48:10 +08:00
BUILD CMSIS-NN: Add transpose conv operator (#2304) 2026-04-20 14:47:53 +08:00
Makefile.inc Port BATCH_MATMUL (#2147) 2023-09-06 16:16:09 +00:00
README.md Sync from upstream TF. (#70) 2021-05-12 07:39:29 -07:00
activation_utils.h Manually sync'd code from upstream TF. (#2) 2021-04-09 11:17:58 -07:00
activations.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
activations.h REF_CODE_REFACTOR: activations.cc (#299) 2021-07-26 23:52:29 +00:00
activations_common.cc Revert "Fix incorrect casting to avoid precision loss (#1501)" (#1536) 2023-02-16 20:04:16 +00:00
activations_test.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
add.cc Initialize ArithmeticParams & OpDataReduce (#2337) 2026-04-20 14:47:55 +08:00
add.h TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
add_common.cc Int32 add (#1847) 2023-03-30 21:39:06 +00:00
add_n.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
add_n_test.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
add_test.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
arg_min_max.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
arg_min_max_test.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
assign_variable.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
batch_matmul.cc Unique Init, Prepare, Eval functions Kernels A-M (#2344) 2026-04-20 14:47:55 +08:00
batch_matmul_test.cc Port BATCH_MATMUL (#2147) 2023-09-06 16:16:09 +00:00
batch_to_space_nd.cc Revert "Compute output shapes for some kernels (#2356)" (#2390) 2026-04-20 14:47:58 +08:00
batch_to_space_nd_test.cc Revert "Compute output shapes for some kernels (#2356)" (#2390) 2026-04-20 14:47:58 +08:00
broadcast_args.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
broadcast_args_test.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
broadcast_to.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
broadcast_to_test.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
call_once.cc Unique Init, Prepare, Eval functions Kernels A-M (#2344) 2026-04-20 14:47:55 +08:00
call_once_test.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
cast.cc Cast compile 2026-04-20 14:48:06 +08:00
cast_test.cc Enable bool input for CAST (#2401) 2026-04-20 14:47:59 +08:00
ceil.cc Unique Init, Prepare, Eval functions Kernels A-M (#2344) 2026-04-20 14:47:55 +08:00
ceil_test.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
circular_buffer.cc Move kTfLiteAbort to micro_context.h (#2072) 2023-07-14 17:10:16 +00:00
circular_buffer.h Move kTfLiteAbort to micro_context.h (#2072) 2023-07-14 17:10:16 +00:00
circular_buffer_common.cc Move kTfLiteAbort to micro_context.h (#2072) 2023-07-14 17:10:16 +00:00
circular_buffer_flexbuffers_generated_data.cc Manually sync'd code from upstream TF. (#2) 2021-04-09 11:17:58 -07:00
circular_buffer_flexbuffers_generated_data.h Manually sync'd code from upstream TF. (#2) 2021-04-09 11:17:58 -07:00
circular_buffer_test.cc Move kTfLiteAbort to micro_context.h (#2072) 2023-07-14 17:10:16 +00:00
comparisons.cc Unique Init, Prepare, Eval functions Kernels A-M (#2344) 2026-04-20 14:47:55 +08:00
comparisons_test.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
concatenation.cc tflite-micro: add concatenation operator compile function 2026-04-20 14:48:09 +08:00
concatenation_test.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
conv.cc Support compiling pure-software conv 2d int8 2026-04-20 14:48:07 +08:00
conv.h tflm: support compile beco on sim. 2026-04-20 14:48:07 +08:00
conv_common.cc Revert "Compute output shapes for some kernels (#2356)" (#2390) 2026-04-20 14:47:58 +08:00
conv_test.cc Revert "Compute output shapes for some kernels (#2356)" (#2390) 2026-04-20 14:47:58 +08:00
conv_test.h Optimized kernel reference fallbacks: CONV (#1911) 2023-08-24 08:30:37 +00:00
conv_test_common.cc Revert "Compute output shapes for some kernels (#2356)" (#2390) 2026-04-20 14:47:58 +08:00
cumsum.cc Unique Init, Prepare, Eval functions Kernels A-M (#2344) 2026-04-20 14:47:55 +08:00
cumsum_test.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
depth_to_space.cc Unique Init, Prepare, Eval functions Kernels A-M (#2344) 2026-04-20 14:47:55 +08:00
depth_to_space_test.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
depthwise_conv.cc Unique Init, Prepare, Eval functions Kernels A-M (#2344) 2026-04-20 14:47:55 +08:00
depthwise_conv.h Revert "Compute output shapes for some kernels (#2356)" (#2390) 2026-04-20 14:47:58 +08:00
depthwise_conv_common.cc Revert "Compute output shapes for some kernels (#2356)" (#2390) 2026-04-20 14:47:58 +08:00
depthwise_conv_test.cc Revert "Compute output shapes for some kernels (#2356)" (#2390) 2026-04-20 14:47:58 +08:00
dequantize.cc Compile dequantize. 2026-04-20 14:48:04 +08:00
dequantize.h dequantize int8 2026-04-20 14:48:02 +08:00
dequantize_common.cc Remove dead code from Dequantize (#2011) 2023-05-31 02:01:17 +00:00
dequantize_test.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
detection_postprocess.cc Unique Init, Prepare, Eval functions Kernels A-M (#2344) 2026-04-20 14:47:55 +08:00
detection_postprocess_flexbuffers_generated_data.cc fix alignment bug in DetectionPostProcess for target based on Arm Corstone-300 software (#1207) 2022-06-27 22:51:34 +00:00
detection_postprocess_flexbuffers_generated_data.h Manually sync'd code from upstream TF. (#2) 2021-04-09 11:17:58 -07:00
detection_postprocess_test.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
div.cc Unique Init, Prepare, Eval functions Kernels A-M (#2344) 2026-04-20 14:47:55 +08:00
div_test.cc Add INT32 support to DIV (#2256) 2023-10-02 18:59:57 +00:00
elementwise.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
elementwise_test.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
elu.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
elu_test.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
embedding_lookup.cc Unique Init, Prepare, Eval functions Kernels A-M (#2344) 2026-04-20 14:47:55 +08:00
embedding_lookup_test.cc Port EMBEDDING_LOOKUP to TFLM (#1872) 2023-05-29 01:25:28 +00:00
ethosu.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
ethosu.h TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
exp.cc Unique Init, Prepare, Eval functions Kernels A-M (#2344) 2026-04-20 14:47:55 +08:00
exp_test.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
expand_dims.cc Expand_dims compile 2026-04-20 14:48:06 +08:00
expand_dims_test.cc Revert "Compute output shapes for some kernels (#2356)" (#2390) 2026-04-20 14:47:58 +08:00
fill.cc Unique Init, Prepare, Eval functions Kernels A-M (#2344) 2026-04-20 14:47:55 +08:00
fill_test.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
floor.cc Unique Init, Prepare, Eval functions Kernels A-M (#2344) 2026-04-20 14:47:55 +08:00
floor_div.cc Unique Init, Prepare, Eval functions Kernels A-M (#2344) 2026-04-20 14:47:55 +08:00
floor_div_test.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
floor_mod.cc Unique Init, Prepare, Eval functions Kernels A-M (#2344) 2026-04-20 14:47:55 +08:00
floor_mod_test.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
floor_test.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
fully_connected.cc Tflm: Add per-channel quantization support to FULLY_CONNECTED operator 2026-04-20 14:48:10 +08:00
fully_connected.h Tflm: Add per-channel quantization support to FULLY_CONNECTED operator 2026-04-20 14:48:10 +08:00
fully_connected_common.cc Tflm: Add per-channel quantization support to FULLY_CONNECTED operator 2026-04-20 14:48:10 +08:00
fully_connected_test.cc Automated sync from github.com/tensorflow/tensorflow (#2221) 2023-09-14 23:18:01 +00:00
gather.cc Unique Init, Prepare, Eval functions Kernels A-M (#2344) 2026-04-20 14:47:55 +08:00
gather_nd.cc Unique Init, Prepare, Eval functions Kernels A-M (#2344) 2026-04-20 14:47:55 +08:00
gather_nd_test.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
gather_test.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
hard_swish.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
hard_swish.h REF_CODE_REFACTOR: hard_swish (#307) 2021-07-27 21:09:42 +00:00
hard_swish_common.cc Replace (almost) all kernel's GetInput/GetOutput with AllocateTempXXx and DeallocateTempXxx micro context api (#894) 2022-02-01 17:53:16 +00:00
hard_swish_test.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
if.cc Unique Init, Prepare, Eval functions Kernels A-M (#2344) 2026-04-20 14:47:55 +08:00
if_test.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
kernel_runner.cc Port BATCH_MATMUL (#2147) 2023-09-06 16:16:09 +00:00
kernel_runner.h Port BATCH_MATMUL (#2147) 2023-09-06 16:16:09 +00:00
kernel_util.cc Compile mechanism base. 2026-04-20 14:48:03 +08:00
kernel_util.h Compile mechanism base. 2026-04-20 14:48:03 +08:00
l2_pool_2d.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
l2_pool_2d_test.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
l2norm.cc Unique Init, Prepare, Eval functions Kernels A-M (#2344) 2026-04-20 14:47:55 +08:00
l2norm_test.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
leaky_relu.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
leaky_relu.h REF_CODE_REFACTOR: Leaky_Relu (#362) 2021-08-17 00:15:05 +00:00
leaky_relu_common.cc Revert "Fix casting to avoid precision loss when mixing double and floats" (#1537) 2023-02-14 00:32:57 +00:00
leaky_relu_test.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
log_softmax.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
log_softmax_test.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
logical.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
logical.h REF_CODE_REFACTOR: logicalAND and logicalOR (#300) 2021-07-27 20:45:41 +00:00
logical_common.cc REF_CODE_REFACTOR: logicalAND and logicalOR (#300) 2021-07-27 20:45:41 +00:00
logical_test.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
logistic.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
logistic.h REF_CODE_REFACTOR: logistic (#308) 2021-07-27 20:01:43 +00:00
logistic_common.cc Input multiplier range check for tanh/logistic kernel. (#2023) 2023-06-21 06:21:17 +00:00
logistic_test.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
lstm_eval.cc Add CMSIS-NN LSTM kernel support (#1620) 2023-03-02 20:44:53 +00:00
lstm_eval.h Add CMSIS-NN LSTM kernel support (#1620) 2023-03-02 20:44:53 +00:00
lstm_eval_common.cc remove lstm internal (#1903) 2023-04-16 06:00:36 +00:00
lstm_eval_test.cc Xtensa LSTM: (#2150) 2026-04-20 14:47:57 +08:00
lstm_eval_test.h Remove the LSTM Internal implementation (#1734) 2023-02-13 19:14:36 +00:00
lstm_shared.h Redesigned reference LSTM kernels (#1682) 2023-02-02 05:43:28 +00:00
maximum_minimum.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
maximum_minimum_test.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
micro_ops.h Port BATCH_MATMUL (#2147) 2023-09-06 16:16:09 +00:00
micro_tensor_utils.cc remove micro_tensor_utils test (#1379) 2022-09-23 22:25:10 +00:00
micro_tensor_utils.h remove micro_tensor_utils test (#1379) 2022-09-23 22:25:10 +00:00
mirror_pad.cc Unique Init, Prepare, Eval functions Kernels A-M (#2344) 2026-04-20 14:47:55 +08:00
mirror_pad_test.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
mul.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
mul.h TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
mul_common.cc Fix missing MUL/int16 quantization case (#1733) 2023-02-14 06:43:28 +00:00
mul_test.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
neg.cc Unique Init, Prepare, Eval functions Kernels N-Z (#2351) 2026-04-20 14:47:56 +08:00
neg_test.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
pack.cc Unique Init, Prepare, Eval functions Kernels N-Z (#2351) 2026-04-20 14:47:56 +08:00
pack_test.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
pad.cc Unique Init, Prepare, Eval functions Kernels N-Z (#2351) 2026-04-20 14:47:56 +08:00
pad.h Allow applications to register custom PAD kernels 2022-11-30 10:28:47 +00:00
pad_test.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
pooling.cc Support compiling pure-software max pool int8. 2026-04-20 14:48:07 +08:00
pooling.h TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
pooling_common.cc Add ref and cmsis-nn kernel support for int16 pooling (#1591) 2023-01-17 18:09:31 +00:00
pooling_test.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
prelu.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
prelu.h Add required header files to prelu.h and dequantize.h (#577) 2021-10-07 02:13:55 +00:00
prelu_common.cc Replace (almost) all kernel's GetInput/GetOutput with AllocateTempXXx and DeallocateTempXxx micro context api (#894) 2022-02-01 17:53:16 +00:00
prelu_test.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
quantization_util_test.cc Manually sync'd code from upstream TF. (#2) 2021-04-09 11:17:58 -07:00
quantize.cc Compile quantize. 2026-04-20 14:48:04 +08:00
quantize.h Compile quantize. 2026-04-20 14:48:04 +08:00
quantize_common.cc Compile quantize. 2026-04-20 14:48:04 +08:00
quantize_test.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
read_variable.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
reduce.cc Compile mean 2026-04-20 14:48:03 +08:00
reduce.h tflm: update tflm xtensa operators codes 2026-04-20 14:48:08 +08:00
reduce_common.cc tflm: update tflm xtensa operators codes 2026-04-20 14:48:08 +08:00
reduce_test.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
reshape.cc Compile reshape 2026-04-20 14:48:04 +08:00
reshape.h Revert "Compute output shapes for some kernels (#2356)" (#2390) 2026-04-20 14:47:58 +08:00
reshape_common.cc Revert "Compute output shapes for some kernels (#2356)" (#2390) 2026-04-20 14:47:58 +08:00
reshape_test.cc Revert "Compute output shapes for some kernels (#2356)" (#2390) 2026-04-20 14:47:58 +08:00
resize_bilinear.cc Unique Init, Prepare, Eval functions Kernels N-Z (#2351) 2026-04-20 14:47:56 +08:00
resize_bilinear_test.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
resize_nearest_neighbor.cc Unique Init, Prepare, Eval functions Kernels N-Z (#2351) 2026-04-20 14:47:56 +08:00
resize_nearest_neighbor_test.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
round.cc Unique Init, Prepare, Eval functions Kernels N-Z (#2351) 2026-04-20 14:47:56 +08:00
round_test.cc Flatten kernel namespace: ROUND (#2113) 2023-07-11 17:34:50 +00:00
select.cc Update SELECT_V2 to match TFLM kernel code (#2015) 2023-06-01 18:42:25 +00:00
select_test.cc Fix SELECT_V2 Prepare deallocation error (#2013) 2023-06-01 03:28:11 +00:00
shape.cc tflite-micro: fix some typos and bugs on specific product 2026-04-20 14:48:08 +08:00
shape_test.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
slice.cc Unique Init, Prepare, Eval functions Kernels N-Z (#2351) 2026-04-20 14:47:56 +08:00
slice_test.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
softmax.cc Support compiling pure-software softmax int8. 2026-04-20 14:48:08 +08:00
softmax.h TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
softmax_common.cc Replace deprecated gen_lut usage by LUTPopulate 2022-11-09 19:31:17 +00:00
softmax_test.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
space_to_batch_nd.cc Revert "Compute output shapes for some kernels (#2356)" (#2390) 2026-04-20 14:47:58 +08:00
space_to_batch_nd_test.cc Revert "Compute output shapes for some kernels (#2356)" (#2390) 2026-04-20 14:47:58 +08:00
space_to_depth.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
space_to_depth_test.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
split.cc Unique Init, Prepare, Eval functions Kernels N-Z (#2351) 2026-04-20 14:47:56 +08:00
split_test.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
split_v.cc Unique Init, Prepare, Eval functions Kernels N-Z (#2351) 2026-04-20 14:47:56 +08:00
split_v_test.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
squared_difference.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
squared_difference_test.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
squeeze.cc Unique Init, Prepare, Eval functions Kernels N-Z (#2351) 2026-04-20 14:47:56 +08:00
squeeze_test.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
strided_slice.cc StridedSlice compile func 2026-04-20 14:48:06 +08:00
strided_slice.h Deduplicate StridedSlice Init and Prepare (#2203) 2023-08-31 23:00:17 +00:00
strided_slice_common.cc Deduplicate StridedSlice Init and Prepare (#2203) 2023-08-31 23:00:17 +00:00
strided_slice_test.cc Manual sync from upstream (and additional changes) (#2018) 2023-06-01 22:41:30 +00:00
sub.cc Sub int compile 2026-04-20 14:48:06 +08:00
sub.h Sub int8 2026-04-20 14:48:06 +08:00
sub_common.cc Revert "Fix incorrect casting to avoid precision loss (#1501)" (#1536) 2023-02-16 20:04:16 +00:00
sub_test.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
svdf.cc Unique Init, Prepare, Eval functions Kernels N-Z (#2351) 2026-04-20 14:47:56 +08:00
svdf.h TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
svdf_common.cc Optimized kernel reference fallbacks: SVDF (#1759) 2023-04-27 23:46:33 +00:00
svdf_test.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
tanh.cc Input multiplier range check for tanh/logistic kernel. (#2023) 2023-06-21 06:21:17 +00:00
tanh_test.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
transpose.cc Transpose compile func 2026-04-20 14:48:05 +08:00
transpose_conv.cc Unique Init, Prepare, Eval functions Kernels N-Z (#2351) 2026-04-20 14:47:56 +08:00
transpose_conv.h CMSIS-NN: Add transpose conv operator (#2304) 2026-04-20 14:47:53 +08:00
transpose_conv_test.cc Automated sync from github.com/tensorflow/tensorflow (#2221) 2023-09-14 23:18:01 +00:00
transpose_test.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
unidirectional_sequence_lstm.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
unidirectional_sequence_lstm.h TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
unidirectional_sequence_lstm_test.cc Update CMSIS-NN CONV and LSTM implementations (#2446) 2026-04-20 14:48:00 +08:00
unpack.cc Unique Init, Prepare, Eval functions Kernels N-Z (#2351) 2026-04-20 14:47:56 +08:00
unpack_test.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00
var_handle.cc Unique Init, Prepare, Eval functions Kernels N-Z (#2351) 2026-04-20 14:47:56 +08:00
while.cc Unique Init, Prepare, Eval functions Kernels N-Z (#2351) 2026-04-20 14:47:56 +08:00
while_test.cc remove usage of AllOpsResolver (#1947) 2023-05-03 02:07:45 +00:00
zeros_like.cc Unique Init, Prepare, Eval functions Kernels N-Z (#2351) 2026-04-20 14:47:56 +08:00
zeros_like_test.cc TFLM swapping TfLiteRegistration_V1 struct to a new more embedded friendly api in the TFLMRegistration struct (#2001) 2023-05-26 06:48:51 +00:00

README.md

Info

  • Porting Ops from Lite to Micro explains, step-by-step, the code changes necessary to port an op from lite to micro, and the process of submitting them for review and acceptance by the project.