Spaces:
Sleeping
Sleeping
Samuel Stevens
commited on
Commit
•
a33c93d
1
Parent(s):
6ee7e7c
try hierarchical averaging
Browse files- app.py +1 -0
- examples/Sarcoscypha-coccinea.jpeg +3 -0
- lib.py +50 -6
- make_txt_embedding.py +48 -5
- test_lib.py +57 -0
app.py
CHANGED
@@ -37,6 +37,7 @@ open_domain_examples = [
|
|
37 |
["examples/Ursus-arctos.jpeg", "Species"],
|
38 |
["examples/Phoca-vitulina.png", "Species"],
|
39 |
["examples/Felis-catus.jpeg", "Genus"],
|
|
|
40 |
]
|
41 |
zero_shot_examples = [
|
42 |
[
|
|
|
37 |
["examples/Ursus-arctos.jpeg", "Species"],
|
38 |
["examples/Phoca-vitulina.png", "Species"],
|
39 |
["examples/Felis-catus.jpeg", "Genus"],
|
40 |
+
["examples/Sarcoscypha-coccinea.jpeg", "Order"],
|
41 |
]
|
42 |
zero_shot_examples = [
|
43 |
[
|
examples/Sarcoscypha-coccinea.jpeg
ADDED
Git LFS Details
|
lib.py
CHANGED
@@ -1,3 +1,13 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
import itertools
|
2 |
import json
|
3 |
|
@@ -33,12 +43,30 @@ class TaxonomicNode:
|
|
33 |
|
34 |
return self._children[first].children(rest)
|
35 |
|
36 |
-
def
|
37 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
38 |
|
39 |
for child in self._children.values():
|
40 |
-
for name, index in child:
|
41 |
-
yield
|
42 |
|
43 |
@classmethod
|
44 |
def from_dict(cls, dct, root):
|
@@ -82,9 +110,25 @@ class TaxonomicTree:
|
|
82 |
|
83 |
return self.kingdoms[first].children(rest)
|
84 |
|
85 |
-
def
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
86 |
for kingdom in self.kingdoms.values():
|
87 |
-
yield from kingdom
|
88 |
|
89 |
def __len__(self):
|
90 |
return self.size
|
|
|
1 |
+
"""
|
2 |
+
Mostly a TaxonomicTree class that implements a taxonomy and some helpers for easily
|
3 |
+
walking and looking in the tree.
|
4 |
+
|
5 |
+
A tree is an arrangement of TaxonomicNodes.
|
6 |
+
|
7 |
+
|
8 |
+
"""
|
9 |
+
|
10 |
+
|
11 |
import itertools
|
12 |
import json
|
13 |
|
|
|
43 |
|
44 |
return self._children[first].children(rest)
|
45 |
|
46 |
+
def descendants(self, prefix=None):
|
47 |
+
"""Iterates over all values in the subtree that match prefix."""
|
48 |
+
|
49 |
+
if not prefix:
|
50 |
+
yield (self.name,), self.index
|
51 |
+
for child in self._children.values():
|
52 |
+
for name, i in child.descendants():
|
53 |
+
yield (self.name, *name), i
|
54 |
+
return
|
55 |
+
|
56 |
+
first, rest = prefix[0], prefix[1:]
|
57 |
+
if first not in self._children:
|
58 |
+
return
|
59 |
+
|
60 |
+
for name, i in self._children[first].descendants(rest):
|
61 |
+
yield (self.name, *name), i
|
62 |
+
|
63 |
+
def values(self):
|
64 |
+
"""Iterates over all (name, i) pairs in the tree."""
|
65 |
+
yield (self.name,), self.index
|
66 |
|
67 |
for child in self._children.values():
|
68 |
+
for name, index in child.values():
|
69 |
+
yield (self.name, *name), index
|
70 |
|
71 |
@classmethod
|
72 |
def from_dict(cls, dct, root):
|
|
|
110 |
|
111 |
return self.kingdoms[first].children(rest)
|
112 |
|
113 |
+
def descendants(self, prefix=None):
|
114 |
+
"""Iterates over all values in the tree that match prefix."""
|
115 |
+
if not prefix:
|
116 |
+
# Give them all the subnodes
|
117 |
+
for kingdom in self.kingdoms.values():
|
118 |
+
yield from kingdom.descendants()
|
119 |
+
|
120 |
+
return
|
121 |
+
|
122 |
+
first, rest = prefix[0], prefix[1:]
|
123 |
+
if first not in self.kingdoms:
|
124 |
+
return
|
125 |
+
|
126 |
+
yield from self.kingdoms[first].descendants(rest)
|
127 |
+
|
128 |
+
def values(self):
|
129 |
+
"""Iterates over all (name, i) pairs in the tree."""
|
130 |
for kingdom in self.kingdoms.values():
|
131 |
+
yield from kingdom.values()
|
132 |
|
133 |
def __len__(self):
|
134 |
return self.size
|
make_txt_embedding.py
CHANGED
@@ -6,20 +6,28 @@ import argparse
|
|
6 |
import csv
|
7 |
import json
|
8 |
import os
|
|
|
9 |
|
10 |
import numpy as np
|
11 |
import torch
|
12 |
import torch.nn.functional as F
|
|
|
13 |
from open_clip import create_model, get_tokenizer
|
14 |
from tqdm import tqdm
|
15 |
|
16 |
import lib
|
17 |
from templates import openai_imagenet_template
|
18 |
|
|
|
|
|
|
|
|
|
19 |
model_str = "hf-hub:imageomics/bioclip"
|
20 |
tokenizer_str = "ViT-B-16"
|
21 |
device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
|
22 |
|
|
|
|
|
23 |
|
24 |
@torch.no_grad()
|
25 |
def write_txt_features(name_lookup):
|
@@ -38,7 +46,7 @@ def write_txt_features(name_lookup):
|
|
38 |
):
|
39 |
# Skip if any non-zero elements
|
40 |
if all_features[:, indices].any():
|
41 |
-
|
42 |
continue
|
43 |
|
44 |
txts = [
|
@@ -59,6 +67,41 @@ def write_txt_features(name_lookup):
|
|
59 |
np.save(args.out_path, all_features)
|
60 |
|
61 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
62 |
def get_name_lookup(catalog_path, cache_path):
|
63 |
if os.path.isfile(cache_path):
|
64 |
with open(cache_path) as fd:
|
@@ -106,14 +149,14 @@ if __name__ == "__main__":
|
|
106 |
args = parser.parse_args()
|
107 |
|
108 |
name_lookup = get_name_lookup(args.catalog_path, cache_path=args.name_cache_path)
|
109 |
-
|
110 |
|
111 |
model = create_model(model_str, output_dict=True, require_pretrained=True)
|
112 |
model = model.to(device)
|
113 |
-
|
114 |
-
|
115 |
model = torch.compile(model)
|
116 |
-
|
117 |
|
118 |
tokenizer = get_tokenizer(tokenizer_str)
|
119 |
write_txt_features(name_lookup)
|
|
|
|
6 |
import csv
|
7 |
import json
|
8 |
import os
|
9 |
+
import logging
|
10 |
|
11 |
import numpy as np
|
12 |
import torch
|
13 |
import torch.nn.functional as F
|
14 |
+
|
15 |
from open_clip import create_model, get_tokenizer
|
16 |
from tqdm import tqdm
|
17 |
|
18 |
import lib
|
19 |
from templates import openai_imagenet_template
|
20 |
|
21 |
+
log_format = "[%(asctime)s] [%(levelname)s] [%(name)s] %(message)s"
|
22 |
+
logging.basicConfig(level=logging.INFO, format=log_format)
|
23 |
+
logger = logging.getLogger()
|
24 |
+
|
25 |
model_str = "hf-hub:imageomics/bioclip"
|
26 |
tokenizer_str = "ViT-B-16"
|
27 |
device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
|
28 |
|
29 |
+
ranks = ("Kingdom", "Phylum", "Class", "Order", "Family", "Genus", "Species")
|
30 |
+
|
31 |
|
32 |
@torch.no_grad()
|
33 |
def write_txt_features(name_lookup):
|
|
|
46 |
):
|
47 |
# Skip if any non-zero elements
|
48 |
if all_features[:, indices].any():
|
49 |
+
logger.info(f"Skipping batch {batch}")
|
50 |
continue
|
51 |
|
52 |
txts = [
|
|
|
67 |
np.save(args.out_path, all_features)
|
68 |
|
69 |
|
70 |
+
def convert_txt_features_to_avgs(name_lookup):
|
71 |
+
assert os.path.isfile(args.out_path)
|
72 |
+
|
73 |
+
# Put that big boy on the GPU. We're going fast.
|
74 |
+
all_features = torch.from_numpy(np.load(args.out_path)).to(device)
|
75 |
+
logger.info("Loaded text features from disk to %s.", device)
|
76 |
+
|
77 |
+
all_names = [set() for rank in ranks]
|
78 |
+
for name, index in tqdm(name_lookup.values()):
|
79 |
+
i = len(name) - 1
|
80 |
+
all_names[i].add((name, index))
|
81 |
+
|
82 |
+
zeroed = 0
|
83 |
+
for i, rank in reversed(list(enumerate(ranks))):
|
84 |
+
if rank == "Species":
|
85 |
+
continue
|
86 |
+
for name, index in tqdm(all_names[i], desc=rank):
|
87 |
+
species = tuple(zip(*((d, i) for d, i in name_lookup.descendants(prefix=name) if len(d) >= 7)))
|
88 |
+
if not species:
|
89 |
+
logger.warning("No species for %s.", " ".join(name))
|
90 |
+
all_features[:, index] = 0.0
|
91 |
+
zeroed += 1
|
92 |
+
continue
|
93 |
+
|
94 |
+
|
95 |
+
values, indices = species
|
96 |
+
mean = all_features[:, indices].mean(dim=1)
|
97 |
+
all_features[:, index] = F.normalize(mean, dim=0)
|
98 |
+
|
99 |
+
out_path, ext = os.path.splitext(args.out_path)
|
100 |
+
np.save(f"{out_path}_avgs{ext}", all_features.cpu().numpy())
|
101 |
+
if zeroed:
|
102 |
+
logger.warning("Zeroed out %d nodes because they didn't have any genus or species-level labels.", zeroed)
|
103 |
+
|
104 |
+
|
105 |
def get_name_lookup(catalog_path, cache_path):
|
106 |
if os.path.isfile(cache_path):
|
107 |
with open(cache_path) as fd:
|
|
|
149 |
args = parser.parse_args()
|
150 |
|
151 |
name_lookup = get_name_lookup(args.catalog_path, cache_path=args.name_cache_path)
|
152 |
+
logger.info("Got name lookup.")
|
153 |
|
154 |
model = create_model(model_str, output_dict=True, require_pretrained=True)
|
155 |
model = model.to(device)
|
156 |
+
logger.info("Created model.")
|
|
|
157 |
model = torch.compile(model)
|
158 |
+
logger.info("Compiled model.")
|
159 |
|
160 |
tokenizer = get_tokenizer(tokenizer_str)
|
161 |
write_txt_features(name_lookup)
|
162 |
+
convert_txt_features_to_avgs(name_lookup)
|
test_lib.py
CHANGED
@@ -422,3 +422,60 @@ def test_taxonomiclookup_children_of_gorilla():
|
|
422 |
)
|
423 |
expected = set()
|
424 |
assert actual == expected
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
422 |
)
|
423 |
expected = set()
|
424 |
assert actual == expected
|
425 |
+
|
426 |
+
|
427 |
+
def test_taxonomictree_descendants_last():
|
428 |
+
lookup = lib.TaxonomicTree()
|
429 |
+
|
430 |
+
lookup.add(("A", "B", "C", "D", "E", "F", "G"))
|
431 |
+
|
432 |
+
actual = list(lookup.descendants(("A", "B", "C", "D", "E", "F", "G")))
|
433 |
+
|
434 |
+
expected = [
|
435 |
+
(("A", "B", "C", "D", "E", "F", "G"), 6),
|
436 |
+
]
|
437 |
+
assert actual == expected
|
438 |
+
|
439 |
+
|
440 |
+
def test_taxonomictree_descendants_entire_tree():
|
441 |
+
lookup = lib.TaxonomicTree()
|
442 |
+
|
443 |
+
lookup.add(("A", "B"))
|
444 |
+
|
445 |
+
actual = list(lookup.descendants())
|
446 |
+
|
447 |
+
expected = [
|
448 |
+
(("A",), 0),
|
449 |
+
(("A", "B"), 1),
|
450 |
+
]
|
451 |
+
assert actual == expected
|
452 |
+
|
453 |
+
|
454 |
+
def test_taxonomictree_descendants_entire_tree_with_prefix():
|
455 |
+
lookup = lib.TaxonomicTree()
|
456 |
+
|
457 |
+
lookup.add(("A", "B"))
|
458 |
+
|
459 |
+
actual = list(lookup.descendants(prefix=("A",)))
|
460 |
+
|
461 |
+
expected = [
|
462 |
+
(("A",), 0),
|
463 |
+
(("A", "B"), 1),
|
464 |
+
]
|
465 |
+
assert actual == expected
|
466 |
+
|
467 |
+
|
468 |
+
def test_taxonomictree_descendants_general():
|
469 |
+
lookup = lib.TaxonomicTree()
|
470 |
+
|
471 |
+
lookup.add(("A", "B", "C", "D", "E", "F", "G"))
|
472 |
+
|
473 |
+
actual = list(lookup.descendants(("A", "B", "C", "D")))
|
474 |
+
|
475 |
+
expected = [
|
476 |
+
(("A", "B", "C", "D"), 3),
|
477 |
+
(("A", "B", "C", "D", "E"), 4),
|
478 |
+
(("A", "B", "C", "D", "E", "F"), 5),
|
479 |
+
(("A", "B", "C", "D", "E", "F", "G"), 6),
|
480 |
+
]
|
481 |
+
assert actual == expected
|