We read every piece of feedback, and take your input very seriously.
To see all available qualifiers, see our documentation.
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
We have calculated the loss of the gate, but does this have any effect on training? Where is this Loss used?
logits = self.wg(input) #dim: [bxs, num_experts] if self.k == 1: self.loss, self.gate_log, gates1_s, dispatch_mask, retval = top1gating( logits, self.capacity_factor if self.training else self.eval_capacity_factor, is_expert_slicing=self.is_expert_slicing, fp16_mode=self.fp16_mode, nonpadding=nonpadding, logits_gumbel=self.logits_gumbel if self.training else 0, token_drop_type=self.token_drop_type, straight_through=self.straight_through, straight_through_temperature=self.straight_through_temperature, balance_ratio=self.balance_ratio, gate_log_req=self.gate_log_req, lid=lid, tutel_cumsum_sub_one=self.tutel_cumsum_sub_one, ) return gates1_s, dispatch_mask, retval
The text was updated successfully, but these errors were encountered:
No branches or pull requests
We have calculated the loss of the gate, but does this have any effect on training? Where is this Loss used?
The text was updated successfully, but these errors were encountered: