Dataset Viewer
Auto-converted to Parquet Duplicate
stage
string
torch
string
cuda_device
string
falcon
string
repo_commit
string
forward_signature
string
dit_path_exists
bool
dit_path_bytes
int64
MXFP8QuantizedLinear_module
null
mxfp8_mod_resolved
bool
mxfp8_mod_has_scaled_mm
bool
mxfp8_mod_has_quant
bool
target_dtype_passed
string
module_type_hist_top
dict
count_MXFP8QuantizedLinear
int64
count_plain_nn_Linear
int64
param_dtype_hist
dict
buffer_dtype_hist
unknown
buffer_examples_by_dtype
unknown
has_float8_e4m3_tensors
bool
has_e8m0_scale_buffers
bool
calls_during_submodule_drive
dict
call_input_dtype_samples
dict
drive_info
dict
verdict
string
done
2.9.0+cu128
NVIDIA B200
0.7.28+torch2.9.1.cu128
e536ded2b3475aa761c48390bb1f866268a590f3
(self, img: torch.Tensor, context: torch.Tensor, t: torch.Tensor, pos: torch.Tensor, mask: torch.Tensor | None = None) -> torch.Tensor
true
13,222,213,290
null
false
false
false
torch.bfloat16
{ "Linear": 264, "RMSNorm": 130, "Attention": 32, "QKNorm": 32, "SwiGLU": 32, "SingleStreamBlock": 28, "DoubleSharedModulation": 28, "TextFusionBlock": 4, "ModuleList": 3, "Sequential": 3, "GELU": 3, "SingleStreamDiT": 1, "PositionalEncoding": 1, "TextFusionTransformer": 1, "LastLayer": 1 ...
0
264
{ "torch.bfloat16": 430 }
{}
{}
false
false
{ "mxfp8_scaled_mm": 0, "quant_mxfp8_triton": 0, "torch_scaled_mm": 0 }
{ "mxfp8_scaled_mm": [], "quant_mxfp8_triton": [], "torch_scaled_mm": [] }
{ "driven": [ { "name": "first", "type": "Linear", "in_features": 64, "out_dtype": "torch.bfloat16" }, { "name": "blocks.0.mlp.down", "type": "Linear", "in_features": 16384, "out_dtype": "torch.bfloat16" }, { "name": "blocks.1.mlp.down", ...
M1

No dataset card yet

Downloads last month
36