File size: 3,174 Bytes
60c8e7f |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 |
pretrained_paths = dict(
BEATs_PATH="/mnt/petrelfs/lixinhao/lxh_exp/pretrained_models/beats/BEATs_iter3+.pt",
UMT_S1_B_PATH="/mnt/lustre/share/videointern/annotations/pretained_model/clipmae_vit_b16_k710_e200.pth",
UMT_S1_L_PATH="/mnt/lustre/share/videointern/annotations/pretained_model/clipmae_vit_l16_k710_e200.pth",
UMT_S1_g_PATH='/mnt/petrelfs/share_data/likunchang/model/um_teacher/umt2/vit_g14_1.1M_CLIP+MAE_300e_pt_k710_ft.pth',
InternVL_6B_PATH = "/mnt/petrelfs/share_data/wangwenhai/internvl/6b_vit_exp126_clip_alpaca_7b_laion5b_peak_1e-5_256gpu_all_trainable_degradation.sh/1499/mp_rank_00_model_states.pt"
)
VisionEncoders = dict()
VisionEncoders["beit"] = dict(
name="beit_base",
pretrained="microsoft/beit-base-patch16-224-pt22k-ft22k",
d_model=768,
)
VisionEncoders["beit_large"] = dict(
name="beit_large",
pretrained="microsoft/beit-large-patch16-224-pt22k-ft22k",
d_model=1024,
)
TextEncoders = dict()
TextEncoders["bert"] = dict(
name="bert_base",
pretrained="bert-base-uncased",
config="configs/config_bert.json",
d_model=768,
fusion_layer=9,
)
TextEncoders["bert_fusion6"] = dict(
name="bert_base_fusion6",
pretrained="bert-base-uncased",
config="configs/config_bert_fusion6.json",
d_model=768,
fusion_layer=6,
)
TextEncoders["bert_large"] = dict(
name="bert_large",
pretrained="bert-large-uncased",
config="configs/config_bert_large.json",
d_model=1024,
fusion_layer=19,
)
TextEncoders["med_bert"] = dict(
name="med_bert_base",
pretrained="bert-base-uncased",
config="configs/med_config.json",
d_model=768,
)
TextEncoders["med_bert_freq2"] = dict(
name="med_bert_base_freq2",
pretrained="bert-base-uncased",
config="configs/med_config_freq2.json",
d_model=768,
)
TextEncoders["med_bert_freq2_must"] = dict(
name="med_bert_base_freq2_must",
pretrained="bert-base-uncased",
config="configs/med_config_freq2_must.json",
d_model=768,
)
TextEncoders["med_bert_fusion10"] = dict(
name="med_bert_base_fusion",
pretrained="bert-base-uncased",
config="configs/med_config_fusion.json",
d_model=768,
fusion_layer=10
)
TextEncoders["med_bert_fusion9"] = dict(
name="med_bert_base_fusion",
pretrained="bert-base-uncased",
config="configs/med_config_fusion.json",
d_model=768,
fusion_layer=9
)
TextEncoders["med_bert_fusion6"] = dict(
name="med_bert_base_fusion",
pretrained="bert-base-uncased",
config="configs/med_config_fusion.json",
d_model=768,
fusion_layer=6
)
TextEncoders["med_bert_fusion0"] = dict(
name="med_bert_base_fusion",
pretrained="bert-base-uncased",
config="configs/med_config_fusion.json",
d_model=768,
fusion_layer=0
)
TextEncoders["med_bert_fusion3"] = dict(
name="med_bert_base_fusion",
pretrained="bert-base-uncased",
config="configs/med_config_fusion.json",
d_model=768,
fusion_layer=3
)
TextEncoders["med_bert_large"] = dict(
name="med_bert_large",
pretrained="bert-base-uncased", # not a bug, it just follows BLIP.
config="configs/med_large_config.json",
d_model=768
) |