import os os.environ["TOKENIZERS_PARALLELISM"] = "false" import logging from tqdm import tqdm from einops import rearrange from transformers.cache_utils import Cache import torch import torch.nn as nn import torch.nn.functional as F import torch.nn.utils.parametrize as P from torch.nn.utils.parametrizations import weight_norm from transformers import LlamaModel, LlamaConfig class LlamaMLP(nn.Module): def __init__(self, hidden_size, intermediate_size): super().__init__() self.hidden_size = hidden_size self.intermediate_size = intermediate_size self.gate_proj = nn.Linear(self.hidden_size, self.intermediate_size, bias=False) self.up_proj = nn.Linear(self.hidden_size, self.intermediate_size, bias=False) self.down_proj = nn.Linear(self.intermediate_size, self.hidden_size, bias=False) self.act_fn = F.silu def forward(self, x): down_proj = self.down_proj(self.act_fn(self.gate_proj(x)) * self.up_proj(x)) return down_proj class GPT_warpper(nn.Module): def __init__( self, gpt_config, num_audio_tokens, num_text_tokens, num_vq=4, **kwargs, ): super().__init__() self.logger = logging.getLogger(__name__) self.gpt = self.build_model(gpt_config) self.model_dim = self.gpt.config.hidden_size self.num_vq = num_vq self.emb_code = nn.ModuleList([nn.Embedding(num_audio_tokens, self.model_dim) for i in range(self.num_vq)]) self.emb_text = nn.Embedding(num_text_tokens, self.model_dim) self.head_text = weight_norm(nn.Linear(self.model_dim, num_text_tokens, bias=False), name='weight') self.head_code = nn.ModuleList([weight_norm(nn.Linear(self.model_dim, num_audio_tokens, bias=False), name='weight') for i in range(self.num_vq)]) def build_model(self, config): configuration = LlamaConfig(**config) model = LlamaModel(configuration) del model.embed_tokens return model def get_emb(self, input_ids, text_mask, **kwargs): emb_text = self.emb_text(input_ids[text_mask][:, 0]) emb_code = [self.emb_code[i](input_ids[~text_mask][:, i]) for i in range(self.num_vq)] emb_code = torch.stack(emb_code, 2).sum(2) emb = torch.zeros((input_ids.shape[:-1])+(emb_text.shape[-1],), device=emb_text.device, dtype=emb_text.dtype) emb[text_mask] = emb_text emb[~text_mask] = emb_code.to(emb.dtype) return emb def prepare_inputs_for_generation( self, input_ids, past_key_values=None, attention_mask=None, inputs_embeds=None, cache_position=None, **kwargs ): # With static cache, the `past_key_values` is None # TODO joao: standardize interface for the different Cache classes and remove of this if has_static_cache = False if past_key_values is None: past_key_values = getattr(self.gpt.layers[0].self_attn, "past_key_value", None) has_static_cache = past_key_values is not None past_length = 0 if past_key_values is not None: if isinstance(past_key_values, Cache): past_length = cache_position[0] if cache_position is not None else past_key_values.get_seq_length() max_cache_length = ( torch.tensor(past_key_values.get_max_length(), device=input_ids.device) if past_key_values.get_max_length() is not None else None ) cache_length = past_length if max_cache_length is None else torch.min(max_cache_length, past_length) # TODO joao: remove this `else` after `generate` prioritizes `Cache` objects else: cache_length = past_length = past_key_values[0][0].shape[2] max_cache_length = None # Keep only the unprocessed tokens: # 1 - If the length of the attention_mask exceeds the length of input_ids, then we are in a setting where # some of the inputs are exclusively passed as part of the cache (e.g. when passing input_embeds as # input) if attention_mask is not None and attention_mask.shape[1] > input_ids.shape[1]: input_ids = input_ids[:, -(attention_mask.shape[1] - past_length) :] # 2 - If the past_length is smaller than input_ids', then input_ids holds all input tokens. We can discard # input_ids based on the past_length. elif past_length < input_ids.shape[1]: input_ids = input_ids[:, past_length:] # 3 - Otherwise (past_length >= input_ids.shape[1]), let's assume input_ids only has unprocessed tokens. # If we are about to go beyond the maximum cache length, we need to crop the input attention mask. if ( max_cache_length is not None and attention_mask is not None and cache_length + input_ids.shape[1] > max_cache_length ): attention_mask = attention_mask[:, -max_cache_length:] position_ids = kwargs.get("position_ids", None) if attention_mask is not None and position_ids is None: # create position_ids on the fly for batch generation position_ids = attention_mask.long().cumsum(-1) - 1 position_ids.masked_fill_(attention_mask == 0, 1) if past_key_values: position_ids = position_ids[:, -input_ids.shape[1] :] # if `inputs_embeds` are passed, we only want to use them in the 1st generation step if inputs_embeds is not None and past_key_values is None: model_inputs = {"inputs_embeds": inputs_embeds} else: # The `contiguous()` here is necessary to have a static stride during decoding. torchdynamo otherwise # recompiles graphs as the stride of the inputs is a guard. Ref: https://github.com/huggingface/transformers/pull/29114 # TODO: use `next_tokens` directly instead. model_inputs = {"input_ids": input_ids.contiguous()} input_length = position_ids.shape[-1] if position_ids is not None else input_ids.shape[-1] if cache_position is None: cache_position = torch.arange(past_length, past_length + input_length, device=input_ids.device) else: cache_position = cache_position[-input_length:] if has_static_cache: past_key_values = None model_inputs.update( { "position_ids": position_ids, "cache_position": cache_position, "past_key_values": past_key_values, "use_cache": kwargs.get("use_cache"), "attention_mask": attention_mask, } ) return model_inputs def generate( self, emb, inputs_ids, temperature, eos_token, attention_mask = None, max_new_token = 2048, min_new_token = 0, LogitsWarpers = [], LogitsProcessors = [], infer_text=False, return_attn=False, return_hidden=False, ): with torch.no_grad(): attentions = [] hiddens = [] start_idx, end_idx = inputs_ids.shape[1], torch.zeros(inputs_ids.shape[0], device=inputs_ids.device, dtype=torch.long) finish = torch.zeros(inputs_ids.shape[0], device=inputs_ids.device).bool() temperature = temperature[None].expand(inputs_ids.shape[0], -1) temperature = rearrange(temperature, "b n -> (b n) 1") attention_mask_cache = torch.ones((inputs_ids.shape[0], inputs_ids.shape[1]+max_new_token,), dtype=torch.bool, device=inputs_ids.device) if attention_mask is not None: attention_mask_cache[:, :attention_mask.shape[1]] = attention_mask for i in tqdm(range(max_new_token)): model_input = self.prepare_inputs_for_generation(inputs_ids, outputs.past_key_values if i!=0 else None, attention_mask_cache[:, :inputs_ids.shape[1]], use_cache=True) if i == 0: model_input['inputs_embeds'] = emb else: if infer_text: model_input['inputs_embeds'] = self.emb_text(model_input['input_ids'][:,:,0]) else: code_emb = [self.emb_code[i](model_input['input_ids'][:,:,i]) for i in range(self.num_vq)] model_input['inputs_embeds'] = torch.stack(code_emb, 3).sum(3) model_input['input_ids'] = None outputs = self.gpt.forward(**model_input, output_attentions=return_attn) attentions.append(outputs.attentions) hidden_states = outputs[0] # 🐻 if return_hidden: hiddens.append(hidden_states[:, -1]) with P.cached(): if infer_text: logits = self.head_text(hidden_states) else: logits = torch.stack([self.head_code[i](hidden_states) for i in range(self.num_vq)], 3) logits = logits[:, -1].float() if not infer_text: logits = rearrange(logits, "b c n -> (b n) c") logits_token = rearrange(inputs_ids[:, start_idx:], "b c n -> (b n) c") else: logits_token = inputs_ids[:, start_idx:, 0] logits = logits / temperature for logitsProcessors in LogitsProcessors: logits = logitsProcessors(logits_token, logits) for logitsWarpers in LogitsWarpers: logits = logitsWarpers(logits_token, logits) if i < min_new_token: logits[:, eos_token] = -torch.inf scores = F.softmax(logits, dim=-1) idx_next = torch.multinomial(scores, num_samples=1) if not infer_text: idx_next = rearrange(idx_next, "(b n) 1 -> b n", n=self.num_vq) finish = finish | (idx_next == eos_token).any(1) inputs_ids = torch.cat([inputs_ids, idx_next.unsqueeze(1)], 1) else: finish = finish | (idx_next == eos_token).any(1) inputs_ids = torch.cat([inputs_ids, idx_next.unsqueeze(-1).expand(-1, -1, self.num_vq)], 1) end_idx = end_idx + (~finish).int() if finish.all(): break inputs_ids = [inputs_ids[idx, start_idx: start_idx+i] for idx, i in enumerate(end_idx.int())] inputs_ids = [i[:, 0] for i in inputs_ids] if infer_text else inputs_ids if return_hidden: hiddens = torch.stack(hiddens, 1) hiddens = [hiddens[idx, :i] for idx, i in enumerate(end_idx.int())] if not finish.all(): self.logger.warn(f'Incomplete result. hit max_new_token: {max_new_token}') return { 'ids': inputs_ids, 'attentions': attentions, 'hiddens':hiddens, }