Commit 3d718bfc authored by rprenger's avatar rprenger
Browse files

Fixing merge conflict

parent 9939fb58

megatron/text_generation_utils.py

+1 −17
Original line number Diff line number Diff line
@@ -175,21 +175,11 @@ def synced_generate(model, context_tokens_tensor, context_length_tensor, max_len 
    if tokens is not None:

        return tokens[:, :context_length], output_logits, full_logits 



<<<<<<< HEAD

def generate(model, sentences=None, max_len=0, all_probs=False):

    if torch.distributed.get_rank() == 0:

        context_tokens_tensor, context_length_tensor = tokenize_batch(sentences)

        c = context_length_tensor[0]

        b = context_tokens_tensor.size(0)

        start = time.time()

        send_generate_info(context_tokens_tensor, context_length_tensor, max_len, all_probs)

=======

def generate(model, sentences=None, max_len=0):

    model.eval()

    if torch.distributed.get_rank() == 0:

        context_tokens_tensor, context_length_tensor = tokenize_batch(sentences)

        send_generate_info(context_tokens_tensor, context_length_tensor, max_len)

>>>>>>> server

        send_generate_info(context_tokens_tensor, context_length_tensor, max_len, all_probs)

    else:

        context_length_tensor, context_tokens_tensor, max_len, all_probs = receive_generate_info()
@@ -206,7 +196,6 @@ def generate(model, sentences=None, max_len=0): 
        decode_tokens = decode_tokens.cpu().numpy().tolist()

        for decode_token in decode_tokens:

            resp_sentences.append(tokenizer.detokenize(decode_token))

<<<<<<< HEAD

            words = []

            for token in decode_token:

                word = tokenizer.tokenizer.decoder[token]
@@ -218,12 +207,7 @@ def generate(model, sentences=None, max_len=0): 
        if all_probs:

            full_logits = full_logits.cpu().numpy().tolist()



        end = time.time()

        print(str(b)+","+str(c)+","+str(len(decode_tokens[0]))+","+str(end-start), flush=True)

        return resp_sentences, resp_sentences_seg, output_logits, full_logits, decode_tokens 

=======

        return resp_sentences

>>>>>>> server



def generate_samples_eval(model, context, max_gen_length, eos_token_id):

    """