You can deploy this model and start earning money today!

Reza2kn/
sapiens2-pretrain-0.8b-INT4-G128

facebook/sapiens2-pretrain-0.8b

quantized

other

Views

No views yet

sapiens2-pretrain-0.8b INT4-G128

Packed 4-bit derivative of facebook/sapiens2-pretrain-0.8b.

This artifact uses symmetric per-group INT4 packing with group size 128 for large floating-point weight tensors. Norms, biases, positional/rope tensors, and small tensors are kept in their source dtype. It is a storage/runtime-loader quant for the current official Sapiens2 code path, not an AWQ/GGUF/NVFP4 LLM artifact.

Files

facebook__sapiens2-pretrain-0.8b-int4-g128.safetensors: packed INT4 safetensors artifact.
load_sapiens2_int4.py: loader that reconstructs a PyTorch state dict for the official Sapiens2 model code.
config.json and preprocessor_config.json: copied from the source repo.
quantization_report.json: build report.

Quantization Report

Source revision: dcc5ac89fcfd01257a68fa580a56b3dda48b4d5f
Group size: 128
Source bytes: 3257134160
Artifact bytes: 422108400
Compression ratio: 7.7163x
Tensors: 550
Quantized tensors: 194
Max tensor MAE during dequant smoke: 0.00414401

Fidelity Validation

The packed INT4 artifact was dequantized back to floating-point tensors and compared against the source checkpoint.

Validation gate: global floating-tensor similarity >= 90.00%
Result: PASS
Global floating-tensor similarity: 99.401621%
Minimum large-tensor cosine: 0.974222064

Loading

python

from load_sapiens2_int4 import load_state_dict

state_dict = load_state_dict("facebook__sapiens2-pretrain-0.8b-int4-g128.safetensors", device="cpu")
# Then instantiate the matching official Sapiens2 architecture and load:
# model.load_state_dict(state_dict, strict=True)

Limitations

This is a verified packed-weight artifact with a dequantizing loader. It does not claim native INT4 CUDA kernels for Sapiens2 yet. Runtime speedups require a Sapiens2-specific kernel/export path and should be benchmarked separately.

Reza2kn/sapiens2-pretrain-0.8b-INT4-G128