mirror of
				https://github.com/ggml-org/llama.cpp.git
				synced 2025-11-03 09:22:01 +00:00 
			
		
		
		
	fix gemma3-cli
This commit is contained in:
		@@ -93,7 +93,7 @@ static int eval_text(gemma3_context & ctx, std::string input, bool logits_last =
 | 
			
		||||
    llama_batch_ext_clear(ctx.batch.get());
 | 
			
		||||
    for (llama_token & t : tokens) {
 | 
			
		||||
        llama_seq_id seq_id = 0;
 | 
			
		||||
        llama_batch_ext_add_text(ctx.batch.get(), t, 0, &seq_id, 1, false);
 | 
			
		||||
        llama_batch_ext_add_text(ctx.batch.get(), t, ctx.n_past++, &seq_id, 1, false);
 | 
			
		||||
    }
 | 
			
		||||
    if (logits_last) {
 | 
			
		||||
        llama_batch_ext_set_output_last(ctx.batch.get());
 | 
			
		||||
 
 | 
			
		||||
		Reference in New Issue
	
	Block a user