remove unnecessary files and code about old version fully_fusible (#481)

* 1119

* 1120

* 1120.2

* model_path

* remove unnecessary files and pre-committed

* remove unnecessary files and pre-committed

* 1121 remove unnecessary files

* modify rev version

* modify rev version

* modify rev version

* accuracy issues targeted

* test script and modify feature

* return set[str]

* add logfile for test

* filter can get the number of kernels in naive_graph_decomposer

* post extract process feature

* remove unnecessary code blocks and variables

* modify the way of counting kernels used

* modify the way of counting kernels used

* modify script, rename files and variables

* add failure protection and log output when removing directories

* add a script to check fusability of a given model

* add a script to check if a given model is fully fusable

* add a script to check if a given model is fully fusable

* a script to check if a given model is fully fusable

* add a script to check if a given model is fully fusionable

* add a script to find fully fusionable subgraph

* find the biggest fully fusionable subgraph

* get fusible subgraph test

* modify get fully fusible subgraph

* improve fully_fusible_subgraph_extractor.py efficiency

* backup code

* Improve efficiency of test/fully_fusible_subgraph_extractor_test.sh

* fully_fusible_subgraph_extractor test

* delete empty file

* modify input file path

* backup

* code bakup

* code bakup

* code bakup

* remove files about old version full_fusible

* remove unnecessary codes about old version fully_fusible

* remove unnecessary codes

* fix graph_net/torch/count_kernels_util.py

---------

Co-authored-by: Xinqi Li <[email protected]>

Author: roll-away

graph_net/config/get_fusible_subgraph_sample_list.txt

@@ -1 +1,11 @@
 samples/timm/resnet18
+samples/timm/eca_resnext26ts.ch_in1k
+samples/timm/eca_vovnet39b
+samples/timm/efficientnet_b0.ra4_e3600_r224_in1k
+samples/timm/efficientnet_b1.ft_in1k
+samples/timm/efficientnet_b2.ra_in1k
+samples/timm/efficientnet_b3.ra2_in1k
+samples/timm/efficientnet_b4.ra2_in1k
+samples/timm/efficientnet_b5.sw_in12k
+samples/timm/efficientnet_el.ra_in1k
+samples/timm/efficientnet_em.ra2_in1k

graph_net/test/dev_model_list/cumsum_num_kernels_sample_list.txt

@@ -1 +1,2 @@
 samples/transformers-auto-model/dbmdz_electra-large-discriminator-finetuned-conll03-english
+samples/timm/test_byobnet.r160_in1k

graph_net/test/dev_model_list/decomposition_error_tmp_torch_samples_list.txt

@@ -4,7 +4,7 @@ GRAPH_NET_ROOT=$(python3 -c "import graph_net; import os; print(
 os.path.dirname(graph_net.__file__))")
 
 # input model path
-MODEL_NAME=resnet18
+MODEL_NAME=test_byobnet.r160_in1k
 MODEL_PATH_IN_SAMPLES=/timm/$MODEL_NAME
 config_json_str=$(cat <<EOF
 {