Replies: 10 comments 27 replies
-
|
constraint_util.py 当您集成维度泛化到子图生成流程时，这个工具会：
_get_in_tensor_symbolic_shapes.py
dimension_generalizer.py
前面三个在OpNamesExtractor前集成 shape_propagator.py
subgraph_input_producer_indexes_generator.py
subgraph_input_shapes_naive_rewriter.py
第一阶段：维度预处理 |
Beta Was this translation helpful? Give feedback.
-
|
这个已经跑通了 |
Beta Was this translation helpful? Give feedback.
-
|
// 在维度泛化之后，opnames提取分析之前 // 在可融合子图生成之前 // 在可融合子图生成之后，重命名之前 |
Beta Was this translation helpful? Give feedback.
-
|
dimension_generalizer → propagate_shapes 举个例子： 这种风格导致其他人没法和它协作，尤其是我们后续还需要建立维度泛化的同源关系（即都源自于同一样本，比如samples/timm/resnet18）。这显然也会让与propagate_shapes的协作进行不下去。 解决办法重构dimension_generalizer的逻辑，让给定 这样后面所有的流程只需要把$output_dir/0/，$output_dir/1/等看成不同的workspace就行。同时同源关系也特别清楚。 |
Beta Was this translation helpful? Give feedback.
-
|
好滴，然后就是test/shape_propagator_test.sh的产物如图，哪里会用到呢？ |
Beta Was this translation helpful? Give feedback.
-
|
子图形状推导要用。看test/subgraph_input_shapes_naive_rewriter_test.sh。 |
Beta Was this translation helpful? Give feedback.
-
|
子图维度泛化后的输入形状信息不是推导出来的。而是“被赋予”的。现有整图的形状推导信息，然后按整图子图关系（input_producer_indexes）一块一块的“切”出来。 |
Beta Was this translation helpful? Give feedback.
-
对的，找到了他的引用了
|
Beta Was this translation helpful? Give feedback.
-
|
这里还有些数量关系需要梳理清楚。
9这个数字的来源是维度泛化会将一份旧样本产生9份新的样本。 |
Beta Was this translation helpful? Give feedback.
-
|
好的 |
Beta Was this translation helpful? Give feedback.
-
|
就算维度泛化产生了100倍的更多输入组合，我们的子图分解流程和整图子图关系（subgraph_sources.json和subgraph_input_producer_index等信息）还是不会变多的。
|
Beta Was this translation helpful? Give feedback.
-
|
好滴 |
Beta Was this translation helpful? Give feedback.
-
|
Beta Was this translation helpful? Give feedback.
-
|
修改以后运行示例，现在就是output_dir/(0-8)/samples...... |
Beta Was this translation helpful? Give feedback.
-
|
Beta Was this translation helpful? Give feedback.
-
|
修改graph_net/dimension_generalizer.py的_get_to_model_path方法，使用数字索引代替复杂的符号维度命名 |
Beta Was this translation helpful? Give feedback.
-
test/workspace_group_ranges_from_subgraph_sources 少了这个 |
Beta Was this translation helpful? Give feedback.
-
python
function group_subgraph_sources() {
echo ">>> [6] Group ranges from subgraph sources."
python3 -m graph_net.apply_sample_pass \
--model-path-list ${range_decomposed_subgraph_list} \
--sample-pass-file-path "$GRAPH_NET_ROOT/graph_net/sample_pass/group_ranges_from_subgraph_sources.py" \
--sample-pass-class-name GroupRangesFromSubgraphSources \
--sample-pass-config $(base64 -w 0 <<EOF
{
"subgraph_model_path_prefix": "${RANGE_DECOMPOSE_OUTPUT_DIR}",
"output_dir": "/tmp/workspace_group_ranges_from_subgraph_sources"
}
EOF
)
}
function analyze_subgraph_dependences(){
python3 -m graph_net.apply_sample_pass \
--model-path-list ${new_model_list} \
--sample-pass-file-path "$GRAPH_NET_ROOT/graph_net/torch/sample_pass/subgraph_input_producer_indexes_generator.py" \
--sample-pass-class-name SubgraphInputProducerIndexesGenerator \
--sample-pass-config $(base64 -w 0 <<EOF
{
"model_path_prefix": "${new_model_prefix}",
"subgraph_ranges_json_root": "/tmp/workspace_group_ranges_from_subgraph_sources",
"subgraph_ranges_json_file_name": "grouped_ranges_from_subgraph_sources.json",
"subgraph_ranges_json_key": "grouped_ranges_from_subgraph_sources",
"output_dir": "/tmp/workspace_subgraph_input_producer_indexes_generator"
}
EOF
)
} |
Beta Was this translation helpful? Give feedback.
-
|
你可以补充一下。这应该用于辅助子图输入形状泛化 |
Beta Was this translation helpful? Give feedback.
-
|
看函数的输入，analyze_subgraph_dependences是需要group_subgraph_sources的输出的 |
Beta Was this translation helpful? Give feedback.
-
|
main() { } main 所以顺序应该是
|
Beta Was this translation helpful? Give feedback.
-
|
重新根据流程更改sh |
Beta Was this translation helpful? Give feedback.
-
|
整图泛化只是为了形状推导 不会基于泛化后的整图再做子图分解的 |
Beta Was this translation helpful? Give feedback.
-
|
|
Beta Was this translation helpful? Give feedback.
-
|
目前我的测试sh的流程 |
Beta Was this translation helpful? Give feedback.
-
|
子图输入生产者下标生成用到了整图，这里线没画出来 |
Beta Was this translation helpful? Give feedback.
-
function gen_fusible_subgraphs() {
echo ">>> [7] Generate fusible subgraphs for subgraph samples under ${DEVICE_REWRITED_OUTPUT_DIR}."
echo ">>>"
python3 -m graph_net.model_path_handler \
--use-subprocess \
--model-path-list "$device_rewrited_subgraph_list" \
--handler-config $(base64 -w 0 <<EOF
{
"handler_path": "$GRAPH_NET_ROOT/graph_net/torch/sample_pass/cumsum_num_kernels_generator.py",
"handler_class_name": "CumSumNumKernelsGenerator",
"handler_config": {
"output_json_file_name": "cumsum_num_kernels.json",
"model_path_prefix": "${DEVICE_REWRITED_OUTPUT_DIR}",
"output_dir": "$CUMSUM_NUM_KERNELS_DIR",
"device": "cuda",
"resume": ${RESUME}
}
}
EOF
)
python3 -m graph_net.model_path_handler \
--model-path-list "$device_rewrited_subgraph_list" \
--handler-config $(base64 -w 0 <<EOF
{
"handler_path": "$GRAPH_NET_ROOT/graph_net/sample_pass/fusible_subgraph_ranges_generator.py",
"handler_class_name": "FusibleSubgraphRangesGenerator",
"handler_config": {
"model_path_prefix": "$CUMSUM_NUM_KERNELS_DIR",
"input_json_file_name": "cumsum_num_kernels.json",
"output_json_file_name": "fusible_subgraph_ranges.json",
"output_dir": "$FUSIBLE_SUBGRAPH_RANGES_DIR",
"resume": ${RESUME}
}
}
EOF
)
python3 -m graph_net.model_path_handler \
--model-path-list "$device_rewrited_subgraph_list" \
--handler-config $(base64 -w 0 <<EOF
{
"handler_path": "$GRAPH_NET_ROOT/graph_net/torch/sample_pass/subgraph_generator.py",
"handler_class_name": "SubgraphGenerator",
"handler_config": {
"model_path_prefix": "${DEVICE_REWRITED_OUTPUT_DIR}",
"output_dir": "$FUSIBLE_SUBGRAPH_SAMPLES_DIR",
"subgraph_ranges_json_root": "$FUSIBLE_SUBGRAPH_RANGES_DIR",
"subgraph_ranges_json_file_name": "fusible_subgraph_ranges.json",
"device": "cuda",
"resume": ${RESUME}
}
}
EOF
)
}第一个python和第二个python是正确运行的，就是第三个python过不去，但是在/graph_net/tools/fusible_subgraph_generate.sh相同功能的脚本进行测试是没有问题的，唯一不同就是它的处理的图是整图 |
Beta Was this translation helpful? Give feedback.
-
|
按照中午说的，我们是在整图list上进行gen_fusible_subgraphs，所以这里的路径是不是应该为model_list而不是分解以后然后去重的子图模型 |
Beta Was this translation helpful? Give feedback.
-
|
这个是graph_net/sample_pass/group_ranges_from_subgraph_sources.py代码问题 |
Beta Was this translation helpful? Give feedback.
-
|
测试以后发现其实不是modellist的问题，是group子图分析的代码有点问题 |
Beta Was this translation helpful? Give feedback.
-
测试完毕
|
Beta Was this translation helpful? Give feedback.
-
|
graph_net/sample_pass/group_ranges_from_subgraph_sources.py 这个代码的逻辑有问题，我修改了 |
Beta Was this translation helpful? Give feedback.
-
|
graph_net/dimension_generalizer.py 这个代码也修改了 |
Beta Was this translation helpful? Give feedback.
-
文件均正常生成，但是未考量过转化率问题
|
Beta Was this translation helpful? Give feedback.
Uh oh!
There was an error while loading. Please reload this page.
-
目前维度泛化相关的功能砖块有：
grpah_net.json里的dim_gen_pass字段tools/init_input_tensor_constraints.shtools/statisticize_in_tensor_symbolic_shapes.shtools/apply_dim_gen_passes.shtest/shape_propagator_test.shtest/subgraph_input_producer_indexes_generator_test.shtest/subgraph_input_shapes_naive_rewriter_test.shBeta Was this translation helpful? Give feedback.
All reactions