ToB企服应用市场:ToB评测及商务社交产业平台
标题:
[条记] rknn Toolkit1.6=>2.0 API变更注释
[打印本页]
作者:
知者何南
时间:
2024-6-8 12:59
标题:
[条记] rknn Toolkit1.6=>2.0 API变更注释
RKNNToolkit2 API Difference With Toolkit1.6
原始文档出处参见上面的标题。下面会对重要的变更内容做标记。
rknn.config
Toolkit1:
config(batch_size=100, # abandoned
caffe_mean_file=None, # abandoned
dtype='float32', # abandoned
epochs=-1, # abandoned
force_gray=None, # abandoned
input_fitting='scale', # abandoned
input_normalization=None, # abandoned
mean_file=None, # abandoned
model_data_format='zone', # abandoned
model_quantize=None, # abandoned
optimize='Default', # abandoned
quantized_dtype='asymmetric_affine-u8',
quantized_moving_alpha=0.01, # abandoned
quantized_algorithm='normal',
quantized_divergence_nbins=1024, # abandoned
mmse_epoch=3, # abandoned
random_brightness=None, # abandoned
random_contrast=None, # abandoned
random_crop=None, # abandoned
random_flip=None, # abandoned
random_mirror=None, # abandoned
reorder_channel=None, # abandoned
restart=False, # abandoned
samples=-1, # abandoned
need_horizontal_merge=False, # abandoned
deconv_merge=True, # abandoned
conv_mul_merge=True, # abandoned
quantized_hybrid=False, # abandoned
output_optimize=0, # abandoned
remove_tensorflow_output_permute=False, # abandoned
optimization_level=3,
target_platform=None,
mean_values=None,
std_values=None,
channel_mean_value=None, # abandoned
force_builtin_perm=False, # abandoned
do_sparse_network=True, # abandoned
merge_dequant_layer_and_output_node=False, # abandoned
quantize_input_node=False, # abandoned
inputs_scale_range=None) # abandoned
复制代码
Toolkit2:
注意量化八月份吧
config(mean_values=None,
std_values=None,
quantized_dtype='asymmetric_quantized-8',
quantized_algorithm='normal',
quantized_method='channel', # new
target_platform=None,
quant_img_RGB2BGR=False, # new
float_dtype='float16', # new
optimization_level=3,
custom_string=None, # new
remove_weight=False, # new
compress_weight=False, # new
inputs_yuv_fmt=None, # new
single_core_mode=False) # new
复制代码
In addition to the above abandoned/new items, there are other differences:
quantized_dtype:
toolkit1: asymmetric_affine-u8, dynamic_fixed_point-i8, dynamic_fixed_point-i16
toolkit2: asymmetric_quantized-8
quantized_algorithm:
toolkit1: normal(default), mmse, kl_divergence, moving_average
toolkit2: normal(default), mmse
target_platform:
toolkit1: rk1808, rk3399pro, rv1109, rv1126
toolkit2: rk3566, rk3568, rk3588, rk3588s, rv1103, rv1106, rk3562 and newer
复制代码
rknn.load_tensorflow
Toolkit1:
load_tensorflow(tf_pb,
inputs,
input_size_list,
outputs,
predef_file=None, # abandoned
mean_values=None, # abandoned
std_values=None, # abandoned
size_with_batch=None) # abandoned
复制代码
Toolkit2:
load_tensorflow(tf_pb,
inputs,
input_size_list,
outputs)
复制代码
In addition to the above abandoned items, there are other differences:
inputs:
toolkit1: node list (layer name)
toolkit2: node list (operand name)
outputs:
toolkit1: node list (layer name)
toolkit2: node list (operand name)
复制代码
rknn.load_caffe
Toolkit1:
load_caffe(model,
proto, # abandoned
blobs=None)
复制代码
Toolkit2:
load_caffe(model,
blobs=None,
input_name=None) # new
复制代码
rknn.load_keras
Toolkit1:
load_keras(model, convert_engine='Keras')
复制代码
Toolkit2:
Not supported yet
!
复制代码
rknn.load_pytorch
Toolkit1:
load_pytorch(model,
input_size_list=None,
inputs=None, # abandoned
outputs=None, # abandoned
convert_engine='torch') # abandoned
复制代码
Toolkit2:
load_pytorch(model,
input_size_list)
复制代码
rknn.load_mxnet
Toolkit1:
load_mxnet(symbol, params, input_size_list=None)
复制代码
Toolkit2:
Not supported yet
!
复制代码
rknn.build
Toolkit1:
build(do_quantization=True,
dataset='dataset.txt',
pre_compile=False, # abandoned
rknn_batch_size=-1)
复制代码
Toolkit2:
build(do_quantization=True,
dataset='dataset.txt',
rknn_batch_size=-1):
复制代码
rknn.direct_build
Toolkit1:
direct_build(model_input, data_input, model_quantize=None, pre_compile=False)
复制代码
Toolkit2:
Not supported yet
!
复制代码
rknn.hybrid_quantization_step1
Toolkit1:
hybrid_quantization_step1(dataset=None)
复制代码
Toolkit2:
hybrid_quantization_step1(dataset=None,
rknn_batch_size=-1, # new
proposal=False, # new
proposal_dataset_size=1) # new
复制代码
rknn.hybrid_quantization_step2
Toolkit1:
hybrid_quantization_step2(model_input,
data_input,
model_quantization_cfg,
dataset, # abandoned
pre_compile=False) # abandoned
复制代码
Toolkit2:
hybrid_quantization_step2(model_input,
data_input,
model_quantization_cfg)
复制代码
rknn.accuracy_analysis
Toolkit1:
accuracy_analysis(inputs,
output_dir='./snapshot',
calc_qnt_error=True, # abandoned
target=None,
device_id=None,
dump_file_type='tensor') # abandoned
复制代码
Toolkit2:
accuracy_analysis(inputs,
output_dir='./snapshot',
target=None,
device_id=None)
复制代码
rknn.load_rknn
Toolkit1:
load_rknn(path,
load_model_in_npu=False) # abandoned
复制代码
Toolkit2:
load_rknn(path)
复制代码
rknn.export_rknn
Toolkit1:
export_rknn(export_path)
复制代码
Toolkit2:
export_rknn(export_path,
**kwargs) # new
复制代码
rknn.load_firmware
Toolkit1:
load_firmware(fw_dir=None)
复制代码
Toolkit2:
Not supported yet
!
复制代码
rknn.init_runtime
Toolkit1:
init_runtime(target=None,
target_sub_class=None,
device_id=None,
perf_debug=False,
eval_mem=False,
async_mode=False,
rknn2precompile=False) # abandoned
复制代码
Toolkit2:
init_runtime(target=None,
target_sub_class=None,
device_id=None,
perf_debug=False,
eval_mem=False,
async_mode=False)
复制代码
In addition to the above abandoned items, there are other differences:
target:
toolkit1: None(simulator), RK3399Pro, RK1808
toolkit2: None(simulator), RK3566, RK3568, RK3588, RK3562
复制代码
rknn.inference
Toolkit1:
inference(inputs,
data_type=None, # abandoned
data_format=None,
inputs_pass_through=None,
get_frame_id=False)
复制代码
Toolkit2:
inference(inputs,
data_format=None,
inputs_pass_through=None,
get_frame_id=False)
复制代码
rknn.eval_perf
Toolkit1:
eval_perf(inputs=None, # abandoned
data_type=None, # abandoned
data_format=None, # abandoned
is_print=True,
loop_cnt=1) # abandoned
复制代码
Toolkit2:
eval_perf(is_print=True)
复制代码
rknn.export_rknn_precompile_model
Toolkit1:
export_rknn_precompile_model(export_path=None)
复制代码
Toolkit2:
Abandoned
复制代码
rknn.export_rknn_sync_model
Toolkit1:
export_rknn_sync_model(input_model=None, sync_uids=None, output_model=None)
复制代码
Toolkit2:
Abandoned
复制代码
rknn.register_op
Toolkit1:
register_op(op_path)
复制代码
Toolkit2:
Not supported yet
复制代码
rknn.fetch_rknn_model_config
Toolkit1:
fetch_rknn_model_config(model_path)
复制代码
Toolkit2:
Not supported yet
复制代码
rknn.list_support_target_platform
Toolkit1:
list_support_target_platform(rknn_model=None)
复制代码
Toolkit2:
Not supported yet
复制代码
免责声明:如果侵犯了您的权益,请联系站长,我们会及时删除侵权内容,谢谢合作!更多信息从访问主页:qidao123.com:ToB企服之家,中国第一个企服评测及商务社交产业平台。
欢迎光临 ToB企服应用市场:ToB评测及商务社交产业平台 (https://dis.qidao123.com/)
Powered by Discuz! X3.4