diff --git a/inference_object.py b/inference_object.py
index e69de29bb2d1d6434b8b29ae775ad8c2e48c5391..576635af756dbdf36d4a99ecd6edcda5343fda3a 100644
--- a/inference_object.py
+++ b/inference_object.py
@@ -0,0 +1,110 @@
+# -*- encoding: utf-8 -*-
+'''
+@File : inference_cogview.py
+@Time : 2021/10/09 19:41:58
+@Author : Ming Ding
+@Contact : dm18@mail.tsinghua.edu.cn
+'''
+
+# here put the import lib
+import os
+import sys
+import math
+import random
+import torch
+import argparse
+
+from arguments import get_args
+from model.cached_autoregressive_model import CachedAutoregressiveModel
+from model.cached_object_model import CachedObjectModel
+from model.ObjectModel import ObjectModel
+from training import load_checkpoint, initialize_distributed, set_random_seed, prepare_tokenizer
+from tokenization import get_tokenizer
+from generation.sampling_strategies import BaseStrategy
+from generation.autoregressive_sampling import filling_sequence
+from generation.utils import timed_name, save_multiple_images, generate_continually
+
+
+def main(args):
+ initialize_distributed(args)
+ tokenizer = prepare_tokenizer(args)
+ # build model
+ model = CachedObjectModel(args)
+ if args.fp16:
+ model = model.half()
+ model = model.to(args.device)
+ load_checkpoint(model, args)
+ set_random_seed(args.seed)
+ model.eval()
+
+ # define function for each query
+ invalid_slices = [slice(tokenizer.img_tokenizer.num_tokens, None)]
+ strategy = BaseStrategy(invalid_slices,
+ temperature=args.temperature, topk=args.top_k)
+
+ def process(raw_text):
+ if args.with_id:
+ query_id, raw_text = raw_text.split('\t')
+ print('raw text: ', raw_text)
+ raw_text = raw_text.split(' ')
+ objects = raw_text[1:]
+ seq = tokenizer.parse_query(f"[ROI1] {raw_text[0]}", img_size=args.img_size)
+ for i in range(len(objects)//5):
+ seq.append(tokenizer['POS0'])
+ seq.extend(objects[i*5:i*5+4] + args.old_token_num)
+ seq.extend(tokenizer.EncodeAsIds(objects[i*5 + 4]))
+ seq.extend(tokenizer.parse_query('[BASE] [BOI1] [MASK]*1024', img_size=args.img_size))
+
+ if len(seq) > 1271:
+ raise ValueError('text too long.')
+ # calibrate text length
+ txt_len = seq.index(tokenizer['[BASE]'])
+ log_attention_weights = torch.zeros(len(seq), len(seq),
+ device=args.device, dtype=torch.half if args.fp16 else torch.float32)
+ log_attention_weights[txt_len + 2:, 1:txt_len] = 1.8 if txt_len <= 10 else 1.4 # TODO args
+ # generation
+ seq = torch.cuda.LongTensor(seq, device=args.device)
+ mbz = args.max_inference_batch_size
+ assert args.batch_size < mbz or args.batch_size % mbz == 0
+ output_list = []
+ for tim in range(max(args.batch_size // mbz, 1)):
+ output_list.append(
+ filling_sequence(model, seq.clone(),
+ batch_size=min(args.batch_size, mbz),
+ strategy=strategy,
+ log_attention_weights=log_attention_weights
+ )
+ )
+ output_tokens = torch.cat(output_list, dim=0)
+ # decoding
+ imgs, txts = [], []
+ for seq in output_tokens:
+ txt_len = seq.index(tokenizer['[BASE]'])
+ seq = seq[txt_len:]
+ _, decoded_imgs = tokenizer.DecodeIds(seq.tolist())
+ imgs.append(decoded_imgs[-1]) # only the last image (target)
+ # save
+ if args.with_id:
+ full_path = os.path.join(args.output_path, query_id)
+ os.makedirs(full_path, exist_ok=True)
+ save_multiple_images(imgs, full_path, False)
+ else:
+ prefix = raw_text.replace('/', '')[:20]
+ full_path = timed_name(prefix, '.jpg', args.output_path)
+ save_multiple_images(imgs, full_path, True)
+
+ os.makedirs(args.output_path, exist_ok=True)
+ generate_continually(process, args.input_source)
+
+
+if __name__ == "__main__":
+ py_parser = argparse.ArgumentParser(add_help=False)
+ py_parser.add_argument('--full-query', action='store_true')
+ py_parser.add_argument('--img-size', type=int, default=256)
+ ObjectModel.add_model_specific_args(py_parser)
+ known, args_list = py_parser.parse_known_args()
+ args = get_args(args_list)
+ args = argparse.Namespace(**vars(args), **vars(known))
+
+ with torch.no_grad():
+ main(args)
\ No newline at end of file
diff --git a/model/ObjectModel.py b/model/ObjectModel.py
index c236f717da63aa2431c1b7b86e9c19d9355608f2..1000fe1f7594cc44c9cd65e9387e0f03d9b18c37 100644
--- a/model/ObjectModel.py
+++ b/model/ObjectModel.py
@@ -3,37 +3,55 @@ import torch.nn.functional as F
from .base_model import BaseModel
-from .mixins import PositionEmbeddingMixin, AttentionMixin
+from .mixins import PositionEmbeddingMixin, AttentionMixin, WordEmebeddingMixin
from mpu.transformer import split_tensor_along_last_dim
from mpu.local_attention_function import f_similar, f_weighting
from mpu.utils import sqrt
from deepspeed.runtime.activation_checkpointing.checkpointing import get_cuda_rng_tracker
+from mpu.mappings import gather_from_model_parallel_region, copy_to_model_parallel_region
class ObjectModel(BaseModel):
def __init__(self, args, transformer=None):
super().__init__(args, transformer=transformer)
additional_seqlen = args.new_sequence_length - args.max_sequence_length
self.mixins.append(PositionEmbeddingMixin(
- additional_seqlen, args.hidden_size
+ additional_seqlen, args.hidden_size,
+ reinit_slice=slice(-180, None)
+ ))
+ self.mixins.append(WordEmebeddingMixin(
+ args.old_token_num, args.additional_token_num, args.hidden_size
))
self.layout = args.layout
def position_embedding_forward(self, position_ids, *other_tensors):
- position = position_ids[..., :self.layout[1]]
- position_plus = position_ids[..., self.layout[1]:]
+ # breakpoint()
+ position_text = position_ids[..., :self.layout[0]]
+ position_object = position_ids[..., self.layout[0]:self.layout[1]]
+ position_image = position_ids[..., self.layout[1]:]
position_embeddings = torch.cat(
(
- self.transformer.position_embeddings(position),
- self.mixins[0].position_embeddings(position_plus)
+ self.transformer.position_embeddings(position_text),
+ self.mixins[0].position_embeddings(position_object),
+ self.transformer.position_embeddings(position_image)
),
dim=-2
)
return position_embeddings
+ def word_embedding_forward(self, input_ids, *other_tensors):
+ return self.mixins[1].word_embeddings(input_ids)
+
+ def final_forward(self, logits, *other_tensors):
+ logits = copy_to_model_parallel_region(logits)
+ logits = F.linear(logits, self.mixins[1].word_embeddings.weight)
+ return logits
+
@classmethod
def add_model_specific_args(cls, parser):
group = parser.add_argument_group('ObjectModel', 'Object model configurations')
- group.add_argument("--layout", type=str, default='64,1088')
- group.add_argument("--new-sequence-length", type=int, default=5185)
+ group.add_argument("--layout", type=str, default='64,246,1270')
+ group.add_argument("--old-token-num", type=int, default=58219)
+ group.add_argument("--additional-token-num", type=int, default=257)
+ group.add_argument("--new-sequence-length", type=int, default=1271) #1089 + 180 + 2
return parser
\ No newline at end of file
diff --git a/model/cached_object_model.py b/model/cached_object_model.py
index e69de29bb2d1d6434b8b29ae775ad8c2e48c5391..6ef67d76ef63fbc905bc9b7155846066f1b9be99 100644
--- a/model/cached_object_model.py
+++ b/model/cached_object_model.py
@@ -0,0 +1,50 @@
+# -*- encoding: utf-8 -*-
+
+# here put the import lib
+import os
+import sys
+import math
+import random
+import torch
+
+from .base_model import BaseModel
+from .ObjectModel import ObjectModel
+from mpu.transformer import standard_attention, split_tensor_along_last_dim
+
+
+class CachedObjectModel(ObjectModel):
+ def __init__(self, args, transformer=None):
+ super().__init__(args, transformer=transformer)
+ self.log_attention_weights = None
+
+ def attention_forward(self, hidden_states, mask, *other_tensors, layer_id=None):
+ attn_module = self.transformer.layers[layer_id].attention
+ mem = other_tensors[layer_id] if len(other_tensors) > 0 else None
+
+ mixed_raw_layer = attn_module.query_key_value(hidden_states)
+ (mixed_query_layer,
+ mixed_key_layer,
+ mixed_value_layer) = split_tensor_along_last_dim(mixed_raw_layer, 3)
+
+ if mem is not None: # the first time, mem is None
+ b = mixed_key_layer.shape[0] # might change batch_size
+ memk, memv = split_tensor_along_last_dim(mem.expand(b, -1, -1), 2)
+ mixed_key_layer = torch.cat((memk, mixed_key_layer), dim=1)
+ mixed_value_layer = torch.cat((memv, mixed_value_layer), dim=1)
+
+ # same as training
+ query_layer = attn_module._transpose_for_scores(mixed_query_layer)
+ key_layer = attn_module._transpose_for_scores(mixed_key_layer)
+ value_layer = attn_module._transpose_for_scores(mixed_value_layer)
+ context_layer = standard_attention(query_layer, key_layer, value_layer, mask, None,
+ log_attention_weights=self.log_attention_weights)
+
+ context_layer = context_layer.permute(0, 2, 1, 3).contiguous()
+ new_context_layer_shape = context_layer.size()[:-2] + (attn_module.hidden_size_per_partition,)
+ context_layer = context_layer.view(*new_context_layer_shape)
+ output = attn_module.dense(context_layer)
+
+ # new mem this layer
+ new_mem = mixed_raw_layer.detach()[..., -(mixed_raw_layer.shape[-1] // 3 * 2):].contiguous()
+
+ return output, new_mem
diff --git a/model/mixins.py b/model/mixins.py
index b8063fb04001dea35a281af72683ad2b41073d64..b5e12dcce595d17d8cdfee591b8aaab564ac717f 100644
--- a/model/mixins.py
+++ b/model/mixins.py
@@ -69,9 +69,16 @@ class AttentionMixin(BaseMixin):
self.dense[layer_id].bias.data.copy_(old_attention.dense.bias.data)
class WordEmebeddingMixin(BaseMixin):
- def __init__(self, additional_token_num, hidden_size):
+ def __init__(self, old_token_num, additional_token_num, hidden_size,
+ init_method_std=0.02, reinit_slice=slice(-1024, None)
+ ):
super(WordEmebeddingMixin, self).__init__()
- self.word_embeddings = torch.nn.Embedding(additional_token_num, hidden_size)
+ self.reinit_slice = reinit_slice
+ self.word_embeddings = torch.nn.Embedding(old_token_num + additional_token_num, hidden_size)
torch.nn.init.normal_(self.word_embeddings.weight, mean=0.0, std=init_method_std)
+
def reinit(self, transformer, *pre_mixins):
- old_weights = transformer.
\ No newline at end of file
+ old_weights = transformer.word_embedding.weight.data
+ old_len, hidden_size = old_weights.shape
+ assert hidden_size == self.word_embeddings.weight.shape[-1]
+ self.word_embeddings.weight[:old_len].data.view(-1, old_len, hidden_size).copy_(old_weights)
\ No newline at end of file
diff --git a/pretrain_objectModel.py b/pretrain_objectModel.py
index e69de29bb2d1d6434b8b29ae775ad8c2e48c5391..0428db3b72aa3cfa67e9a8f1b7d6771de5d8ea7a 100644
--- a/pretrain_objectModel.py
+++ b/pretrain_objectModel.py
@@ -0,0 +1,207 @@
+# -*- encoding: utf-8 -*-
+'''
+@File : pretrain_cogview2.py
+@Time : 2021/10/06 00:58:32
+@Author : Ming Ding
+@Contact : dm18@mail.tsinghua.edu.cn
+'''
+
+# here put the import lib
+import os
+import sys
+import math
+import random
+import torch
+import argparse
+import numpy as np
+
+import mpu
+from arguments import get_args
+from model.ObjectModel import ObjectModel
+from training.deepspeed_training import training_main
+from data_utils import BinaryDataset
+from tokenization import get_tokenizer
+from tokenization.cogview import TextCodeTemplate
+
+def get_names():
+ return ["背景",
+ "人","自行车","汽车","摩托车","飞机","公交车","火车","卡车","船","红绿灯",
+ "消防栓","空", "停车牌","停车收费表","长椅","鸟","猫","狗","马","羊",
+ "牛","大象","熊","斑马","长颈鹿","空","背包","伞","空","空",
+ "手提包","领带","手提箱","飞盘","滑雪板","滑雪板","球","风筝","棒球棒","棒球手套",
+ "滑板","冲浪板","网球拍","瓶子","空", "酒杯","杯子","叉子","刀子","勺子",
+ "碗","香蕉","苹果","三明治","橘子","花椰菜","胡萝卜","热狗","比萨饼","甜甜圈",
+ "蛋糕","椅子","沙发","盆栽植物","床","空", "餐桌","空","空","厕所",
+ "空","电视","笔记本","鼠标","遥控器","键盘","手机","微波炉","烤箱","烤面包机",
+ "水槽","冰箱","空","书","钟","花瓶","剪刀","泰迪熊","吹风机","牙刷"]
+
+
+def get_masks_and_position_ids(data,
+ n_pads,
+ object_pads,
+ loss_mask=None,
+ attention_mask=None, args=None):
+ # Extract batch size and sequence length.
+ batch_size, seq_length = data.size()
+
+ # Attention mask (lower triangular).
+ if attention_mask is None:
+ assert loss_mask is not None
+ # loss_mask has n_pad(+1 CLS and [1:] then) zeros, so it is the same as attention_mask, reuse.
+ attention_mask = loss_mask[:, :seq_length].unsqueeze(-2).expand(batch_size, seq_length, seq_length).tril()
+ for i in range(batch_size):
+ attention_mask[i].fill_diagonal_(1)
+ attention_mask = attention_mask.unsqueeze(1)
+
+ # Loss mask.
+ if loss_mask is None:
+ loss_mask = torch.ones(data.size(), dtype=data.dtype, device=data.device)
+
+ # Position ids.
+ #1270
+ position_ids = torch.zeros(batch_size, seq_length, dtype=torch.long,
+ device=data.device)
+
+ for i in range(batch_size):
+ torch.arange(64 - n_pads[i], out=position_ids[i, n_pads[i]:64],
+ dtype=torch.long, device=data.device)
+ torch.arange(180-object_pads[i], out=position_ids[i, 64+object_pads:64+180])
+ # breakpoint()
+ torch.arange(64 - n_pads[i], 64 - n_pads[i] + seq_length - (64+180),
+ out=position_ids[i, 64+180:],
+ dtype=torch.long, device=data.device)
+ return attention_mask, loss_mask, position_ids
+
+
+def get_batch(data_iterator, args, timers):
+ # Items and their type.
+ keys = ['text', 'loss_mask', 'object_pad', 'n_pad']
+ datatype = torch.int64
+
+ # Broadcast data.
+ timers('data loader').start()
+ if data_iterator is not None:
+ data = next(data_iterator)
+ else:
+ data = None
+ timers('data loader').stop()
+
+ data_b = mpu.broadcast_data(keys, data, datatype)
+ # Unpack.
+ # breakpoint()
+ tokens_ = data_b['text'].long()
+ loss_mask = data_b['loss_mask'].float()
+ n_pads = data_b['n_pad'].long()
+ object_pads = data_b['object_pad'].long()
+
+ labels = tokens_[:, 1:].contiguous()
+ loss_mask = loss_mask[:, 1:].contiguous()
+ tokens = tokens_[:, :-1].contiguous()
+
+ attention_mask = None
+
+ # Get the masks and postition ids.
+ attention_mask, loss_mask, position_ids = get_masks_and_position_ids(
+ tokens,
+ n_pads,
+ object_pads,
+ loss_mask=loss_mask,
+ attention_mask=attention_mask,
+ args=args
+ )
+ # Convert
+ if args.fp16:
+ attention_mask = attention_mask.half()
+
+ return tokens, labels, loss_mask, attention_mask, position_ids
+
+
+def forward_step(data_iterator, model, args, timers):
+ """Forward step."""
+
+ # Get the batch.
+ timers('batch generator').start()
+ tokens, labels, loss_mask, attention_mask, position_ids = get_batch(
+ data_iterator, args, timers)
+ timers('batch generator').stop()
+ # Forward model.
+ logits, *mems = model(tokens, position_ids, attention_mask)
+ losses = mpu.vocab_parallel_cross_entropy(logits.contiguous().float(), labels)
+ # scaling loss mask
+ loss_mask = loss_mask.view(-1)
+
+ losses = losses.view(-1) * loss_mask
+ loss = torch.sum(losses) / loss_mask.sum()
+
+ return loss, {}
+
+def create_dataset_function(path, args):
+ tokenizer = get_tokenizer()
+ layout = [100,164,1188]
+ names = get_names()
+ tokens = []
+ for name in names:
+ tokens.append(tokenizer.EncodeAsIds(name))
+ # 4 + 4 + 1 一个object需要9个token 20个需要 20*9 = 180个
+ def process_fn(row):
+ row = row.astype(np.int64)
+ codes = [row[layout[1]:layout[2]]]
+ text = row[layout[0]:layout[1]]
+ text = text[text > 0][:63] # [ROI]
+ object_tokens = []
+ # print(row[:100])
+ for i in range(20):
+ object = row[i * 5: (i+1) * 5]
+ if object[0] == -1:
+ break
+ # print("object", object)
+ object[2] += object[0]
+ object[3] += object[1]
+ object_tokens.append(tokenizer['[POS0]'])
+ object_tokens.extend([object[j] + args.old_token_num for j in range(4)])
+ object_tokens.extend(tokens[object[4]])
+ object_tokens = np.array(object_tokens)
+ # print(object_tokens)
+ object_pad = 180 - object_tokens.shape[-1]
+ object_tokens = np.concatenate([
+ np.array([tokenizer['[PAD]']] * object_pad, dtype=np.int64),
+ object_tokens
+ ], axis = 0)
+ # 180
+ text_object = np.concatenate([text, np.array(object_tokens, dtype=np.int64)], axis=0)
+ # print(len(text), len(text_object))
+ # print(len(codes[0]))
+ merged = TextCodeTemplate(text_object, codes[0], tokenizer)
+ # print(len(merged), len(text))
+ n_pad = args.new_sequence_length - len(merged)
+ parts = [
+ np.array([tokenizer['[PAD]']] * n_pad, dtype=np.int64),
+ merged
+ ]
+ ret = np.concatenate(parts, axis=0)
+ return {'text': ret,
+ 'loss_mask': np.array([0] * n_pad + [1] * (len(text) + 1) + [0] * object_pad + [1] * (182 - object_pad + 1025)),
+ 'object_pad':object_pad,
+ 'n_pad':n_pad
+ }
+
+ return BinaryDataset(path, process_fn, length_per_sample=layout[-1])
+
+
+
+if __name__ == '__main__':
+ py_parser = argparse.ArgumentParser(add_help=False)
+
+ py_parser.add_argument('--txt-loss-scale', type=float, default=1)
+
+ ObjectModel.add_model_specific_args(py_parser)
+
+ known, args_list = py_parser.parse_known_args()
+
+ args = get_args(args_list)
+ args = argparse.Namespace(**vars(args), **vars(known))
+
+ args.layout = [int(x) for x in args.layout.split(',')]
+
+ training_main(args, model_cls=ObjectModel, forward_step_function=forward_step,
+ create_dataset_function=create_dataset_function)
diff --git a/scripts/finetune_into_object.sh b/scripts/finetune_into_object.sh
index e69de29bb2d1d6434b8b29ae775ad8c2e48c5391..9cae1aae45e65ec3c86c63804c6e81a617c26ec1 100644
--- a/scripts/finetune_into_object.sh
+++ b/scripts/finetune_into_object.sh
@@ -0,0 +1,60 @@
+#! /bin/bash
+
+# Change for multinode config
+
+NUM_WORKERS=4
+NUM_GPUS_PER_WORKER=8
+MP_SIZE=1
+
+script_path=$(realpath $0)
+script_dir=$(dirname $script_path)
+main_dir=$(dirname $script_dir)
+
+OPTIONS_NCCL="NCCL_DEBUG=info NCCL_IB_DISABLE=0 NCCL_NET_GDR_LEVEL=2"
+HOST_FILE_PATH="hostfile"
+
+full_data='/dataset/fd5061f6/cogview/cogdata_new/cogdata_task_coco_detection_task/coco/coco.bin.part_0.cogdata'
+
+# --mode finetune \
+# --resume-dataloader \
+# --load pretrained/cogview/cogview-base
+config_json="$script_dir/ds_config_zero.json"
+gpt_options=" \
+ --experiment-name finetune2-object-test \
+ --tokenizer-type cogview \
+ --img-tokenizer-path pretrained/vqvae/vqvae_hard_biggerset_011.pt \
+ --model-parallel-size ${MP_SIZE} \
+ --num-layers 48 \
+ --hidden-size 2560 \
+ --num-attention-heads 40 \
+ --train-iters 200000 \
+ --train-data ${full_data} \
+ --split 949,50,1 \
+ --distributed-backend nccl \
+ --lr-decay-style cosine \
+ --warmup .1 \
+ --checkpoint-activations \
+ --max-sequence-length 1089 \
+ --sandwich-ln \
+ --fp16 \
+ --save-interval 2000 \
+ --eval-interval 1000 \
+ --save $main_dir/checkpoints \
+ --mode finetune \
+ --resume-dataloader \
+ --load pretrained/cogview/cogview-base
+"
+ # --load pretrained/cogview/cogview-base
+
+
+gpt_options="${gpt_options}
+ --deepspeed \
+ --deepspeed_config ${config_json} \
+"
+
+
+run_cmd="${OPTIONS_NCCL} deepspeed --num_nodes ${NUM_WORKERS} --num_gpus ${NUM_GPUS_PER_WORKER} --hostfile ${HOST_FILE_PATH} pretrain_objectModel.py $@ ${gpt_options}"
+echo ${run_cmd}
+eval ${run_cmd}
+
+set +x
diff --git a/scripts/kill_python.sh b/scripts/kill_python.sh
index e69de29bb2d1d6434b8b29ae775ad8c2e48c5391..3900363636355cbdd970edaff9d622eef166337a 100644
--- a/scripts/kill_python.sh
+++ b/scripts/kill_python.sh
@@ -0,0 +1 @@
+pdsh -w ssh:node[1-3] "pkill -9 python"
\ No newline at end of file
diff --git a/scripts/pretrain_multiple_nodes.sh b/scripts/pretrain_multiple_nodes.sh
index 5a09f7712528974023124dd510d3762623e7c8d9..78e665fb2ca7c9b872e5708ae2af8b0ae2bf47b2 100755
--- a/scripts/pretrain_multiple_nodes.sh
+++ b/scripts/pretrain_multiple_nodes.sh
@@ -2,7 +2,7 @@
# Change for multinode config
-NUM_WORKERS=1
+NUM_WORKERS=3
NUM_GPUS_PER_WORKER=8
MP_SIZE=1
@@ -12,7 +12,6 @@ main_dir=$(dirname $script_dir)
OPTIONS_NCCL="NCCL_DEBUG=info NCCL_IB_DISABLE=0 NCCL_NET_GDR_LEVEL=2"
HOST_FILE_PATH="hostfile"
-HOST_FILE_PATH="hostfile_single"
full_data="/dataset/fd5061f6/cogview/cogdata_new/cogdata_task_4leveltokens/merge.bin"
small_data="/dataset/fd5061f6/cogview/cogdata_new/cogdata_task_4leveltokens/zijian/zijian.bin.part_0.cogdata"
diff --git a/scripts/text2image_object.sh b/scripts/text2image_object.sh
index 96e8a045bbb2f3d4f046b9949dc734b12ae0288e..4ab4b70b745a3c47dfdd03c1f9ef80645ef531b3 100755
--- a/scripts/text2image_object.sh
+++ b/scripts/text2image_object.sh
@@ -1,6 +1,6 @@
#!/bin/bash
-CHECKPOINT_PATH=pretrained/cogview/cogview2-base
+CHECKPOINT_PATH=checkpoints/finetune2-object-test10-24-08-21
NLAYERS=48
NHIDDEN=2560
NATT=40
@@ -15,9 +15,9 @@ TOPK=200
script_path=$(realpath $0)
script_dir=$(dirname $script_path)
-MASTER_PORT=${MASTER_PORT} python inference_cogview2.py \
+MASTER_PORT=${MASTER_PORT} python inference_object.py \
--tokenizer-type cogview \
- --img-tokenizer-path pretrained/vqvae/l1+ms-ssim+revd_percep.pt \
+ --img-tokenizer-path pretrained/vqvae/vqvae_hard_biggerset_011.pt \
--mode inference \
--distributed-backend nccl \
--max-sequence-length 1089 \
@@ -32,10 +32,10 @@ MASTER_PORT=${MASTER_PORT} python inference_cogview2.py \
--top_k $TOPK \
--sandwich-ln \
--input-source ./input.txt \
- --output-path samples_text2image \
+ --output-path samples_text2image_object \
--batch-size 4 \
--max-inference-batch-size 8 \
- --device 0 \
+ --device 7 \
$@