mirror of https://github.com/google/gemma.cpp.git
Address review feedback: Fix prefill_tbatch_size and variable placement issues
This commit is contained in:
parent
27c28cc938
commit
f55c321397
|
|
@ -156,8 +156,7 @@ void ReplGemma(const ThreadingArgs& threading, const InferenceArgs& inference,
|
|||
std::cout << token_text << std::flush;
|
||||
return true;
|
||||
};
|
||||
// Flag to check if we should exit after processing non-interactive prompt
|
||||
bool exit_after_generation = !inference.prompt.empty();
|
||||
|
||||
while (true) { // Loop until user quits.
|
||||
tokens_generated_this_turn = 0;
|
||||
|
||||
|
|
@ -224,7 +223,7 @@ void ReplGemma(const ThreadingArgs& threading, const InferenceArgs& inference,
|
|||
std::cout << "\n\n";
|
||||
|
||||
// Break the loop if in non-interactive mode
|
||||
if (exit_after_generation) {
|
||||
if (!inference.prompt.empty()) {
|
||||
break;
|
||||
}
|
||||
|
||||
|
|
|
|||
Loading…
Reference in New Issue