config.json Examples
The following examples contain all configurable parameters in all supported scenarios.
PyTorch
task = statistics
{
"task": "statistics",
"dump_path": "/home/data_dump",
"rank": [],
"step": [],
"level": "L1",
"async_dump": false,
"extra_info": true,
"statistics": {
"scope": [],
"list": [],
"tensor_list": [],
"data_mode": ["all"],
"summary_mode": "statistics"
}
}
task = tensor
{
"task": "tensor",
"dump_path": "/home/data_dump",
"rank": [],
"step": [],
"level": "L1",
"async_dump": false,
"tensor": {
"scope": [],
"list":[],
"data_mode": ["all"],
"bench_path": "/home/bench_data_dump",
"summary_mode": "md5",
"diff_nums": 5
}
}
task = acc_check
{
"task": "acc_check",
"dump_path": "/home/data_dump",
"rank": [],
"step": [],
"level": "L1",
"acc_check": {
"white_list": [],
"black_list": [],
"error_data_path": "./"
}
}
task = structure
{
"task": "structure",
"dump_path": "/home/data_dump",
"rank": [],
"step": [],
"level": "mix"
}
Example of Dynamic Control of dump_enable
{
"task": "statistics",
"dump_path": "/home/data_dump",
"rank": [],
"step": [],
"level": "L1",
"dump_enable": false,
"statistics": {
"summary_mode": "statistics"
}
}
Note:
dump_enableis configured only when dump needs to be dynamically enabled or disabled. During execution, you can change the value ofdump_enablefromfalsetotrue(or vice versa) to dynamically enable or disable dump. Modifications to other fields in the JSON file also take effect.
MindSpore Static Graph
task = statistics
{
"task": "statistics",
"dump_path": "/home/data_dump",
"rank": [],
"step": [],
"level": "L2",
"statistics": {
"list": [],
"data_mode": ["all"],
"summary_mode": "statistics"
}
}
task = tensor
{
"task": "tensor",
"dump_path": "/home/data_dump",
"rank": [],
"step": [],
"level": "L2",
"tensor": {
"list":[],
"data_mode": ["all"]
}
}
task = overflow_check
{
"task": "overflow_check",
"dump_path": "/home/data_dump",
"rank": [],
"step": [],
"level": "L2",
"overflow_check": {
"check_mode": "all"
}
}
task = exception_dump
{
"task": "exception_dump",
"dump_path": "/home/data_dump",
"rank": [],
"step": [],
"level": "L2"
}
MindSpore Dynamic Graph
task = statistics
{
"task": "statistics",
"dump_path": "/home/data_dump",
"rank": [],
"step": [],
"level": "L1",
"statistics": {
"scope": [],
"list": [],
"data_mode": ["all"],
"summary_mode": "statistics"
}
}
task = tensor
{
"task": "tensor",
"dump_path": "/home/data_dump",
"rank": [],
"step": [],
"level": "L1",
"tensor": {
"scope": [],
"list":[],
"data_mode": ["all"]
}
}
task = structure
{
"task": "structure",
"dump_path": "/home/data_dump",
"rank": [],
"step": [],
"level": "mix"
}
task = exception_dump
{
"task": "exception_dump",
"dump_path": "/home/data_dump",
"rank": [],
"step": [],
"level": "L2"
}