caixudong2001 commited on
Commit Β·
bd6a9bf
1
Parent(s): 528ef77
Replace checkpoints with inference weights
Browse files- checkpoints/checkpoints7/config.json +0 -92
- checkpoints/checkpoints7/pytorch_model.bin +0 -3
- checkpoints/checkpoints8/special_tokens_map.json +0 -20
- checkpoints/checkpoints8/tokenizer_config.json +0 -62
- checkpoints/checkpoints8/vocab.txt +0 -15
- checkpoints/{checkpoints1 β oriongeno_arthropods}/config.json +0 -0
- checkpoints/{checkpoints5 β oriongeno_arthropods}/pytorch_model.bin +0 -0
- checkpoints/{checkpoints1 β oriongeno_arthropods}/special_tokens_map.json +0 -0
- checkpoints/{checkpoints1 β oriongeno_arthropods}/tokenizer_config.json +0 -0
- checkpoints/{checkpoints1 β oriongeno_arthropods}/vocab.txt +0 -0
- checkpoints/{checkpoints2 β oriongeno_birds}/config.json +0 -0
- checkpoints/{checkpoints3 β oriongeno_birds}/pytorch_model.bin +0 -0
- checkpoints/{checkpoints2 β oriongeno_birds}/special_tokens_map.json +0 -0
- checkpoints/{checkpoints2 β oriongeno_birds}/tokenizer_config.json +0 -0
- checkpoints/{checkpoints2 β oriongeno_birds}/vocab.txt +0 -0
- checkpoints/{checkpoints3 β oriongeno_fish}/config.json +0 -0
- checkpoints/{checkpoints2 β oriongeno_fish}/pytorch_model.bin +0 -0
- checkpoints/{checkpoints3 β oriongeno_fish}/special_tokens_map.json +0 -0
- checkpoints/{checkpoints3 β oriongeno_fish}/tokenizer_config.json +0 -0
- checkpoints/{checkpoints3 β oriongeno_fish}/vocab.txt +0 -0
- checkpoints/{checkpoints5 β oriongeno_fungi}/config.json +0 -0
- checkpoints/{checkpoints6 β oriongeno_fungi}/pytorch_model.bin +1 -1
- checkpoints/{checkpoints4 β oriongeno_fungi}/special_tokens_map.json +0 -0
- checkpoints/{checkpoints4 β oriongeno_fungi}/tokenizer_config.json +0 -0
- checkpoints/{checkpoints4 β oriongeno_fungi}/vocab.txt +0 -0
- checkpoints/{checkpoints6 β oriongeno_mammals}/config.json +0 -0
- checkpoints/{checkpoints1 β oriongeno_mammals}/pytorch_model.bin +0 -0
- checkpoints/{checkpoints5 β oriongeno_mammals}/special_tokens_map.json +0 -0
- checkpoints/{checkpoints5 β oriongeno_mammals}/tokenizer_config.json +0 -0
- checkpoints/{checkpoints5 β oriongeno_mammals}/vocab.txt +0 -0
- checkpoints/{checkpoints4 β oriongeno_other_vertebrates}/config.json +0 -0
- checkpoints/{checkpoints4 β oriongeno_other_vertebrates}/pytorch_model.bin +0 -0
- checkpoints/{checkpoints6 β oriongeno_other_vertebrates}/special_tokens_map.json +0 -0
- checkpoints/{checkpoints6 β oriongeno_other_vertebrates}/tokenizer_config.json +0 -0
- checkpoints/{checkpoints6 β oriongeno_other_vertebrates}/vocab.txt +0 -0
- checkpoints/{checkpoints8 β oriongeno_plants}/config.json +0 -0
- checkpoints/{checkpoints8 β oriongeno_plants}/pytorch_model.bin +0 -0
- checkpoints/{checkpoints7 β oriongeno_plants}/special_tokens_map.json +0 -0
- checkpoints/{checkpoints7 β oriongeno_plants}/tokenizer_config.json +0 -0
- checkpoints/{checkpoints7 β oriongeno_plants}/vocab.txt +0 -0
checkpoints/checkpoints7/config.json
DELETED
|
@@ -1,92 +0,0 @@
|
|
| 1 |
-
{
|
| 2 |
-
"architectures": [
|
| 3 |
-
"OrionGenoForJointLabeling"
|
| 4 |
-
],
|
| 5 |
-
"attention_type": "flash_attention_2",
|
| 6 |
-
"attn_cfg": {},
|
| 7 |
-
"attn_layer_idx": [],
|
| 8 |
-
"bidirectional": true,
|
| 9 |
-
"bidirectional_strategy": "add",
|
| 10 |
-
"bidirectional_weight_tie": true,
|
| 11 |
-
"bos_token_id": 14,
|
| 12 |
-
"d_intermediate": 0,
|
| 13 |
-
"d_model": 768,
|
| 14 |
-
"dtype": "bfloat16",
|
| 15 |
-
"eos_token_id": 13,
|
| 16 |
-
"fused_add_norm": true,
|
| 17 |
-
"initializer_cfg": null,
|
| 18 |
-
"keys_to_ignore_at_inference": [
|
| 19 |
-
"gene_logits",
|
| 20 |
-
"repeat_logits",
|
| 21 |
-
"boundary_logits",
|
| 22 |
-
"gene_cross_entropy",
|
| 23 |
-
"repeat_cross_entropy",
|
| 24 |
-
"boundary_focal_loss",
|
| 25 |
-
"hidden_states"
|
| 26 |
-
],
|
| 27 |
-
"loss_weights": {
|
| 28 |
-
"boundary_focal_loss": 0.25,
|
| 29 |
-
"repeat_cross_entropy": 0.1,
|
| 30 |
-
"state_cross_entropy": 1.0
|
| 31 |
-
},
|
| 32 |
-
"mask_token_id": 10,
|
| 33 |
-
"model_type": "orion_geno",
|
| 34 |
-
"n_layer": 24,
|
| 35 |
-
"norm_epsilon": 1e-05,
|
| 36 |
-
"pad_token_id": 9,
|
| 37 |
-
"pad_vocab_size_multiple": 1,
|
| 38 |
-
"residual_in_fp32": true,
|
| 39 |
-
"rms_norm": true,
|
| 40 |
-
"settings": {
|
| 41 |
-
"channel_growth": 128,
|
| 42 |
-
"decoder_channel_sizes": [
|
| 43 |
-
768,
|
| 44 |
-
768,
|
| 45 |
-
640,
|
| 46 |
-
512,
|
| 47 |
-
384
|
| 48 |
-
],
|
| 49 |
-
"encoder_output_channels": 384,
|
| 50 |
-
"feature_scales": [
|
| 51 |
-
1,
|
| 52 |
-
2,
|
| 53 |
-
4,
|
| 54 |
-
8
|
| 55 |
-
],
|
| 56 |
-
"gene_label_count": 20,
|
| 57 |
-
"input_channels": 5,
|
| 58 |
-
"loss": {
|
| 59 |
-
"boundary_label_ids": [
|
| 60 |
-
7,
|
| 61 |
-
8,
|
| 62 |
-
9,
|
| 63 |
-
10,
|
| 64 |
-
11,
|
| 65 |
-
12,
|
| 66 |
-
13,
|
| 67 |
-
14,
|
| 68 |
-
17,
|
| 69 |
-
19
|
| 70 |
-
],
|
| 71 |
-
"boundary_loss_gamma": 2.0,
|
| 72 |
-
"boundary_positive_weight": 4.0,
|
| 73 |
-
"intergenic_label_id": 0,
|
| 74 |
-
"state_boundary_weight": 1.6,
|
| 75 |
-
"state_default_weight": 1.0,
|
| 76 |
-
"state_intergenic_weight": 0.25,
|
| 77 |
-
"state_label_smoothing": 0.02,
|
| 78 |
-
"state_near_boundary_distance": 8,
|
| 79 |
-
"state_near_boundary_factor": 1.2
|
| 80 |
-
},
|
| 81 |
-
"pooling_kernel_size": 2,
|
| 82 |
-
"repeat_label_count": 2,
|
| 83 |
-
"use_decoder_skip_connection": false,
|
| 84 |
-
"use_repeat_head": true
|
| 85 |
-
},
|
| 86 |
-
"ssm_cfg": {
|
| 87 |
-
"layer": "Mamba2"
|
| 88 |
-
},
|
| 89 |
-
"transformers_version": "4.57.3",
|
| 90 |
-
"use_second_lm_head": false,
|
| 91 |
-
"vocab_size": 15
|
| 92 |
-
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
checkpoints/checkpoints7/pytorch_model.bin
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:9701780388e4a4c06248d97c193821663b95c8a4c466332acee73174be2c86ec
|
| 3 |
-
size 244045419
|
|
|
|
|
|
|
|
|
|
|
|
checkpoints/checkpoints8/special_tokens_map.json
DELETED
|
@@ -1,20 +0,0 @@
|
|
| 1 |
-
{
|
| 2 |
-
"bos_token": {
|
| 3 |
-
"content": "<bos>",
|
| 4 |
-
"lstrip": false,
|
| 5 |
-
"normalized": false,
|
| 6 |
-
"rstrip": false,
|
| 7 |
-
"single_word": false
|
| 8 |
-
},
|
| 9 |
-
"cls_token": "<cls>",
|
| 10 |
-
"eos_token": {
|
| 11 |
-
"content": "<eos>",
|
| 12 |
-
"lstrip": false,
|
| 13 |
-
"normalized": false,
|
| 14 |
-
"rstrip": false,
|
| 15 |
-
"single_word": false
|
| 16 |
-
},
|
| 17 |
-
"mask_token": "<mask>",
|
| 18 |
-
"pad_token": "<pad>",
|
| 19 |
-
"unk_token": "<unk>"
|
| 20 |
-
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
checkpoints/checkpoints8/tokenizer_config.json
DELETED
|
@@ -1,62 +0,0 @@
|
|
| 1 |
-
{
|
| 2 |
-
"added_tokens_decoder": {
|
| 3 |
-
"9": {
|
| 4 |
-
"content": "<pad>",
|
| 5 |
-
"lstrip": false,
|
| 6 |
-
"normalized": false,
|
| 7 |
-
"rstrip": false,
|
| 8 |
-
"single_word": false,
|
| 9 |
-
"special": true
|
| 10 |
-
},
|
| 11 |
-
"10": {
|
| 12 |
-
"content": "<mask>",
|
| 13 |
-
"lstrip": false,
|
| 14 |
-
"normalized": false,
|
| 15 |
-
"rstrip": false,
|
| 16 |
-
"single_word": false,
|
| 17 |
-
"special": true
|
| 18 |
-
},
|
| 19 |
-
"11": {
|
| 20 |
-
"content": "<unk>",
|
| 21 |
-
"lstrip": false,
|
| 22 |
-
"normalized": false,
|
| 23 |
-
"rstrip": false,
|
| 24 |
-
"single_word": false,
|
| 25 |
-
"special": true
|
| 26 |
-
},
|
| 27 |
-
"12": {
|
| 28 |
-
"content": "<cls>",
|
| 29 |
-
"lstrip": false,
|
| 30 |
-
"normalized": false,
|
| 31 |
-
"rstrip": false,
|
| 32 |
-
"single_word": false,
|
| 33 |
-
"special": true
|
| 34 |
-
},
|
| 35 |
-
"13": {
|
| 36 |
-
"content": "<eos>",
|
| 37 |
-
"lstrip": false,
|
| 38 |
-
"normalized": false,
|
| 39 |
-
"rstrip": false,
|
| 40 |
-
"single_word": false,
|
| 41 |
-
"special": true
|
| 42 |
-
},
|
| 43 |
-
"14": {
|
| 44 |
-
"content": "<bos>",
|
| 45 |
-
"lstrip": false,
|
| 46 |
-
"normalized": false,
|
| 47 |
-
"rstrip": false,
|
| 48 |
-
"single_word": false,
|
| 49 |
-
"special": true
|
| 50 |
-
}
|
| 51 |
-
},
|
| 52 |
-
"bos_token": "<bos>",
|
| 53 |
-
"clean_up_tokenization_spaces": true,
|
| 54 |
-
"cls_token": "<cls>",
|
| 55 |
-
"eos_token": "<eos>",
|
| 56 |
-
"extra_special_tokens": {},
|
| 57 |
-
"mask_token": "<mask>",
|
| 58 |
-
"model_max_length": 51200,
|
| 59 |
-
"pad_token": "<pad>",
|
| 60 |
-
"tokenizer_class": "DnaTokenizer",
|
| 61 |
-
"unk_token": "<unk>"
|
| 62 |
-
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
checkpoints/checkpoints8/vocab.txt
DELETED
|
@@ -1,15 +0,0 @@
|
|
| 1 |
-
A
|
| 2 |
-
C
|
| 3 |
-
G
|
| 4 |
-
T
|
| 5 |
-
N
|
| 6 |
-
a
|
| 7 |
-
c
|
| 8 |
-
g
|
| 9 |
-
t
|
| 10 |
-
<pad>
|
| 11 |
-
<mask>
|
| 12 |
-
<unk>
|
| 13 |
-
<cls>
|
| 14 |
-
<eos>
|
| 15 |
-
<bos>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
checkpoints/{checkpoints1 β oriongeno_arthropods}/config.json
RENAMED
|
File without changes
|
checkpoints/{checkpoints5 β oriongeno_arthropods}/pytorch_model.bin
RENAMED
|
File without changes
|
checkpoints/{checkpoints1 β oriongeno_arthropods}/special_tokens_map.json
RENAMED
|
File without changes
|
checkpoints/{checkpoints1 β oriongeno_arthropods}/tokenizer_config.json
RENAMED
|
File without changes
|
checkpoints/{checkpoints1 β oriongeno_arthropods}/vocab.txt
RENAMED
|
File without changes
|
checkpoints/{checkpoints2 β oriongeno_birds}/config.json
RENAMED
|
File without changes
|
checkpoints/{checkpoints3 β oriongeno_birds}/pytorch_model.bin
RENAMED
|
File without changes
|
checkpoints/{checkpoints2 β oriongeno_birds}/special_tokens_map.json
RENAMED
|
File without changes
|
checkpoints/{checkpoints2 β oriongeno_birds}/tokenizer_config.json
RENAMED
|
File without changes
|
checkpoints/{checkpoints2 β oriongeno_birds}/vocab.txt
RENAMED
|
File without changes
|
checkpoints/{checkpoints3 β oriongeno_fish}/config.json
RENAMED
|
File without changes
|
checkpoints/{checkpoints2 β oriongeno_fish}/pytorch_model.bin
RENAMED
|
File without changes
|
checkpoints/{checkpoints3 β oriongeno_fish}/special_tokens_map.json
RENAMED
|
File without changes
|
checkpoints/{checkpoints3 β oriongeno_fish}/tokenizer_config.json
RENAMED
|
File without changes
|
checkpoints/{checkpoints3 β oriongeno_fish}/vocab.txt
RENAMED
|
File without changes
|
checkpoints/{checkpoints5 β oriongeno_fungi}/config.json
RENAMED
|
File without changes
|
checkpoints/{checkpoints6 β oriongeno_fungi}/pytorch_model.bin
RENAMED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 244045419
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c07455e26f50a21de82b1a27489350c2085d6f69f417bf441fbee098139f8a3e
|
| 3 |
size 244045419
|
checkpoints/{checkpoints4 β oriongeno_fungi}/special_tokens_map.json
RENAMED
|
File without changes
|
checkpoints/{checkpoints4 β oriongeno_fungi}/tokenizer_config.json
RENAMED
|
File without changes
|
checkpoints/{checkpoints4 β oriongeno_fungi}/vocab.txt
RENAMED
|
File without changes
|
checkpoints/{checkpoints6 β oriongeno_mammals}/config.json
RENAMED
|
File without changes
|
checkpoints/{checkpoints1 β oriongeno_mammals}/pytorch_model.bin
RENAMED
|
File without changes
|
checkpoints/{checkpoints5 β oriongeno_mammals}/special_tokens_map.json
RENAMED
|
File without changes
|
checkpoints/{checkpoints5 β oriongeno_mammals}/tokenizer_config.json
RENAMED
|
File without changes
|
checkpoints/{checkpoints5 β oriongeno_mammals}/vocab.txt
RENAMED
|
File without changes
|
checkpoints/{checkpoints4 β oriongeno_other_vertebrates}/config.json
RENAMED
|
File without changes
|
checkpoints/{checkpoints4 β oriongeno_other_vertebrates}/pytorch_model.bin
RENAMED
|
File without changes
|
checkpoints/{checkpoints6 β oriongeno_other_vertebrates}/special_tokens_map.json
RENAMED
|
File without changes
|
checkpoints/{checkpoints6 β oriongeno_other_vertebrates}/tokenizer_config.json
RENAMED
|
File without changes
|
checkpoints/{checkpoints6 β oriongeno_other_vertebrates}/vocab.txt
RENAMED
|
File without changes
|
checkpoints/{checkpoints8 β oriongeno_plants}/config.json
RENAMED
|
File without changes
|
checkpoints/{checkpoints8 β oriongeno_plants}/pytorch_model.bin
RENAMED
|
File without changes
|
checkpoints/{checkpoints7 β oriongeno_plants}/special_tokens_map.json
RENAMED
|
File without changes
|
checkpoints/{checkpoints7 β oriongeno_plants}/tokenizer_config.json
RENAMED
|
File without changes
|
checkpoints/{checkpoints7 β oriongeno_plants}/vocab.txt
RENAMED
|
File without changes
|