Hello Haofei and UniMatch maintainers,
Thank you for releasing GMFlow and UniMatch. Could you clarify which license applies to the pretrained checkpoint gmflow-scale1-things-e9887eda.pth, linked from UniMatch's MODEL_ZOO.md?
We see Apache-2.0 in the original GMFlow repository and MIT in UniMatch. Does either license cover this S3-hosted checkpoint, or are there separate terms for the pretrained weights?
Our immediate intended use is a local inference-only evaluation on 15 image pairs from video material we are authorized to use, without training, redistribution of weights, or deployment. Is this use covered by the released terms?
Separately, if we later consider commercial inference or distributing the weights with software, are there additional requirements or restrictions? A link to the applicable existing terms would be sufficient.
For exact identification, the downloaded file is 18,768,907 bytes with SHA256 e9887eda6165e0813c03f35998d6af86e265aafce2e3349af93d691a59696088.
Thank you.
Hello Haofei and UniMatch maintainers,
Thank you for releasing GMFlow and UniMatch. Could you clarify which license applies to the pretrained checkpoint
gmflow-scale1-things-e9887eda.pth, linked from UniMatch's MODEL_ZOO.md?We see Apache-2.0 in the original GMFlow repository and MIT in UniMatch. Does either license cover this S3-hosted checkpoint, or are there separate terms for the pretrained weights?
Our immediate intended use is a local inference-only evaluation on 15 image pairs from video material we are authorized to use, without training, redistribution of weights, or deployment. Is this use covered by the released terms?
Separately, if we later consider commercial inference or distributing the weights with software, are there additional requirements or restrictions? A link to the applicable existing terms would be sufficient.
For exact identification, the downloaded file is 18,768,907 bytes with SHA256
e9887eda6165e0813c03f35998d6af86e265aafce2e3349af93d691a59696088.Thank you.