tarekziade HF Staff commited on
Commit
47543c8
·
1 Parent(s): 91ff508

Review cloud MIC baselines 7d8449a176b7481f9cb779992b3f8cb3 (#1)

Browse files

- Review cloud MIC baselines 7d8449a176b7481f9cb779992b3f8cb3 (a7e259f584c09e459ce4f0ebf7de2e03413a9ea4)

integrations/transformers/expectations/cpu-x86_64-bf16/detr-mic.bfloat16.json CHANGED
@@ -1,34 +1,44 @@
1
  {
2
  "recorded_with": {
3
- "contract": "hf-internal-testing/mic-detr-integration",
4
- "contract_name": "detr",
5
- "contract_revision": "f3a79837574f4bc72ca3201bb01b85696fe37a0c",
6
  "framework": "transformers",
7
  "variant": "bfloat16",
8
  "target": "detr-mic",
9
  "kind": "fixture",
10
- "revision": "3a3a1420126e33823f073575fdb3bbac07037b52",
11
  "environment": {
12
  "python": "3.11.17",
13
- "transformers": "5.19.0.dev0",
14
- "torch": "2.11.0+cpu",
15
- "torchvision": "0.26.0+cpu",
16
  "torchaudio": "2.11.0+cpu",
17
- "tokenizers": "0.23.1",
18
- "numpy": "1.26.4",
19
- "pillow": "12.2.0",
20
- "accelerate": "1.13.0",
21
  "safetensors": "0.8.0",
22
- "huggingface-hub": "1.32.0",
23
- "sentencepiece": "0.2.1",
24
- "protobuf": "6.33.6",
25
  "librosa": "0.11.0",
26
- "soundfile": "0.13.1",
27
- "scipy": "1.15.3"
28
  },
29
- "seconds": 7.29,
30
  "environment_differs": {
31
- "python": "3.11.17 (pinned 3.10)"
 
 
 
 
 
 
 
 
 
 
32
  },
33
  "platform": "cpu-x86_64-bf16"
34
  },
 
1
  {
2
  "recorded_with": {
3
+ "contract": "huggingface-mic/detr-mic",
4
+ "contract_name": "mic-detr",
5
+ "contract_revision": "91ff508188db759f2ee39ec64016a7e417f13e4d",
6
  "framework": "transformers",
7
  "variant": "bfloat16",
8
  "target": "detr-mic",
9
  "kind": "fixture",
10
+ "revision": "91ff508188db759f2ee39ec64016a7e417f13e4d",
11
  "environment": {
12
  "python": "3.11.17",
13
+ "transformers": "5.21.0.dev0",
14
+ "torch": "2.14.0+cpu",
15
+ "torchvision": "0.29.1+cpu",
16
  "torchaudio": "2.11.0+cpu",
17
+ "tokenizers": "0.23.3",
18
+ "numpy": "2.3.5",
19
+ "pillow": "12.3.0",
20
+ "accelerate": "1.15.0",
21
  "safetensors": "0.8.0",
22
+ "huggingface-hub": "2.3.0",
23
+ "sentencepiece": "0.2.2",
24
+ "protobuf": "7.36.2",
25
  "librosa": "0.11.0",
26
+ "soundfile": "0.14.0",
27
+ "scipy": "1.17.1"
28
  },
29
+ "seconds": 8.6,
30
  "environment_differs": {
31
+ "torch": "2.14.0+cpu (pinned 2.11.0)",
32
+ "torchvision": "0.29.1+cpu (pinned 0.26.0)",
33
+ "tokenizers": "0.23.3 (pinned 0.23.1)",
34
+ "numpy": "2.3.5 (pinned 1.26.4)",
35
+ "pillow": "12.3.0 (pinned 12.2.0)",
36
+ "accelerate": "1.15.0 (pinned 1.13.0)",
37
+ "huggingface-hub": "2.3.0 (pinned 1.32.0)",
38
+ "sentencepiece": "0.2.2 (pinned 0.2.1)",
39
+ "protobuf": "7.36.2 (pinned 6.33.6)",
40
+ "soundfile": "0.14.0 (pinned 0.13.1)",
41
+ "scipy": "1.17.1 (pinned 1.15.3)"
42
  },
43
  "platform": "cpu-x86_64-bf16"
44
  },
integrations/transformers/expectations/cpu-x86_64/detr-mic.eager.json CHANGED
@@ -1,34 +1,44 @@
1
  {
2
  "recorded_with": {
3
- "contract": "hf-internal-testing/mic-detr-integration",
4
- "contract_name": "detr",
5
- "contract_revision": "f3a79837574f4bc72ca3201bb01b85696fe37a0c",
6
  "framework": "transformers",
7
  "variant": "eager",
8
  "target": "detr-mic",
9
  "kind": "fixture",
10
- "revision": "3a3a1420126e33823f073575fdb3bbac07037b52",
11
  "environment": {
12
  "python": "3.11.17",
13
- "transformers": "5.19.0.dev0",
14
- "torch": "2.11.0+cpu",
15
- "torchvision": "0.26.0+cpu",
16
  "torchaudio": "2.11.0+cpu",
17
- "tokenizers": "0.23.1",
18
- "numpy": "1.26.4",
19
- "pillow": "12.2.0",
20
- "accelerate": "1.13.0",
21
  "safetensors": "0.8.0",
22
- "huggingface-hub": "1.32.0",
23
- "sentencepiece": "0.2.1",
24
- "protobuf": "6.33.6",
25
  "librosa": "0.11.0",
26
- "soundfile": "0.13.1",
27
- "scipy": "1.15.3"
28
  },
29
- "seconds": 7.3,
30
  "environment_differs": {
31
- "python": "3.11.17 (pinned 3.10)"
 
 
 
 
 
 
 
 
 
 
32
  },
33
  "platform": "cpu-x86_64"
34
  },
 
1
  {
2
  "recorded_with": {
3
+ "contract": "huggingface-mic/detr-mic",
4
+ "contract_name": "mic-detr",
5
+ "contract_revision": "91ff508188db759f2ee39ec64016a7e417f13e4d",
6
  "framework": "transformers",
7
  "variant": "eager",
8
  "target": "detr-mic",
9
  "kind": "fixture",
10
+ "revision": "91ff508188db759f2ee39ec64016a7e417f13e4d",
11
  "environment": {
12
  "python": "3.11.17",
13
+ "transformers": "5.21.0.dev0",
14
+ "torch": "2.14.0+cpu",
15
+ "torchvision": "0.29.1+cpu",
16
  "torchaudio": "2.11.0+cpu",
17
+ "tokenizers": "0.23.3",
18
+ "numpy": "2.3.5",
19
+ "pillow": "12.3.0",
20
+ "accelerate": "1.15.0",
21
  "safetensors": "0.8.0",
22
+ "huggingface-hub": "2.3.0",
23
+ "sentencepiece": "0.2.2",
24
+ "protobuf": "7.36.2",
25
  "librosa": "0.11.0",
26
+ "soundfile": "0.14.0",
27
+ "scipy": "1.17.1"
28
  },
29
+ "seconds": 9.36,
30
  "environment_differs": {
31
+ "torch": "2.14.0+cpu (pinned 2.11.0)",
32
+ "torchvision": "0.29.1+cpu (pinned 0.26.0)",
33
+ "tokenizers": "0.23.3 (pinned 0.23.1)",
34
+ "numpy": "2.3.5 (pinned 1.26.4)",
35
+ "pillow": "12.3.0 (pinned 12.2.0)",
36
+ "accelerate": "1.15.0 (pinned 1.13.0)",
37
+ "huggingface-hub": "2.3.0 (pinned 1.32.0)",
38
+ "sentencepiece": "0.2.2 (pinned 0.2.1)",
39
+ "protobuf": "7.36.2 (pinned 6.33.6)",
40
+ "soundfile": "0.14.0 (pinned 0.13.1)",
41
+ "scipy": "1.17.1 (pinned 1.15.3)"
42
  },
43
  "platform": "cpu-x86_64"
44
  },
integrations/transformers/expectations/cpu-x86_64/detr-mic.json CHANGED
@@ -1,34 +1,44 @@
1
  {
2
  "recorded_with": {
3
- "contract": "hf-internal-testing/mic-detr-integration",
4
- "contract_name": "detr",
5
- "contract_revision": "f3a79837574f4bc72ca3201bb01b85696fe37a0c",
6
  "framework": "transformers",
7
  "variant": "default",
8
  "target": "detr-mic",
9
  "kind": "fixture",
10
- "revision": "3a3a1420126e33823f073575fdb3bbac07037b52",
11
  "environment": {
12
  "python": "3.11.17",
13
- "transformers": "5.19.0.dev0",
14
- "torch": "2.11.0+cpu",
15
- "torchvision": "0.26.0+cpu",
16
  "torchaudio": "2.11.0+cpu",
17
- "tokenizers": "0.23.1",
18
- "numpy": "1.26.4",
19
- "pillow": "12.2.0",
20
- "accelerate": "1.13.0",
21
  "safetensors": "0.8.0",
22
- "huggingface-hub": "1.32.0",
23
- "sentencepiece": "0.2.1",
24
- "protobuf": "6.33.6",
25
  "librosa": "0.11.0",
26
- "soundfile": "0.13.1",
27
- "scipy": "1.15.3"
28
  },
29
- "seconds": 7.55,
30
  "environment_differs": {
31
- "python": "3.11.17 (pinned 3.10)"
 
 
 
 
 
 
 
 
 
 
32
  },
33
  "platform": "cpu-x86_64"
34
  },
 
1
  {
2
  "recorded_with": {
3
+ "contract": "huggingface-mic/detr-mic",
4
+ "contract_name": "mic-detr",
5
+ "contract_revision": "91ff508188db759f2ee39ec64016a7e417f13e4d",
6
  "framework": "transformers",
7
  "variant": "default",
8
  "target": "detr-mic",
9
  "kind": "fixture",
10
+ "revision": "91ff508188db759f2ee39ec64016a7e417f13e4d",
11
  "environment": {
12
  "python": "3.11.17",
13
+ "transformers": "5.21.0.dev0",
14
+ "torch": "2.14.0+cpu",
15
+ "torchvision": "0.29.1+cpu",
16
  "torchaudio": "2.11.0+cpu",
17
+ "tokenizers": "0.23.3",
18
+ "numpy": "2.3.5",
19
+ "pillow": "12.3.0",
20
+ "accelerate": "1.15.0",
21
  "safetensors": "0.8.0",
22
+ "huggingface-hub": "2.3.0",
23
+ "sentencepiece": "0.2.2",
24
+ "protobuf": "7.36.2",
25
  "librosa": "0.11.0",
26
+ "soundfile": "0.14.0",
27
+ "scipy": "1.17.1"
28
  },
29
+ "seconds": 9.56,
30
  "environment_differs": {
31
+ "torch": "2.14.0+cpu (pinned 2.11.0)",
32
+ "torchvision": "0.29.1+cpu (pinned 0.26.0)",
33
+ "tokenizers": "0.23.3 (pinned 0.23.1)",
34
+ "numpy": "2.3.5 (pinned 1.26.4)",
35
+ "pillow": "12.3.0 (pinned 12.2.0)",
36
+ "accelerate": "1.15.0 (pinned 1.13.0)",
37
+ "huggingface-hub": "2.3.0 (pinned 1.32.0)",
38
+ "sentencepiece": "0.2.2 (pinned 0.2.1)",
39
+ "protobuf": "7.36.2 (pinned 6.33.6)",
40
+ "soundfile": "0.14.0 (pinned 0.13.1)",
41
+ "scipy": "1.17.1 (pinned 1.15.3)"
42
  },
43
  "platform": "cpu-x86_64"
44
  },
integrations/transformers/expectations/cuda/detr-mic.bfloat16.json CHANGED
@@ -1,32 +1,45 @@
1
  {
2
  "recorded_with": {
3
- "contract": "hf-internal-testing/mic-detr-integration",
4
- "contract_name": "detr",
5
- "contract_revision": "unpublished",
6
  "framework": "transformers",
7
  "variant": "bfloat16",
8
  "target": "detr-mic",
9
  "kind": "fixture",
10
- "revision": "unpublished",
11
  "environment": {
12
- "python": "3.10.12",
13
- "transformers": "5.19.0.dev0",
14
- "torch": "2.11.0+cu128",
15
- "torchvision": "0.26.0+cu128",
16
- "torchaudio": "2.11.0+cu128",
17
- "tokenizers": "0.23.1",
18
- "numpy": "1.26.4",
19
- "pillow": "12.2.0",
20
- "accelerate": "1.13.0",
21
  "safetensors": "0.8.0",
22
- "huggingface-hub": "1.32.0",
23
- "sentencepiece": "0.2.1",
24
- "protobuf": "6.33.6",
25
  "librosa": "0.11.0",
26
- "soundfile": "0.13.1",
27
- "scipy": "1.15.3"
 
 
 
 
 
 
 
 
 
 
 
 
 
 
28
  },
29
- "seconds": 5.06,
30
  "platform": "cuda"
31
  },
32
  "outputs": {
 
1
  {
2
  "recorded_with": {
3
+ "contract": "huggingface-mic/detr-mic",
4
+ "contract_name": "mic-detr",
5
+ "contract_revision": "91ff508188db759f2ee39ec64016a7e417f13e4d",
6
  "framework": "transformers",
7
  "variant": "bfloat16",
8
  "target": "detr-mic",
9
  "kind": "fixture",
10
+ "revision": "91ff508188db759f2ee39ec64016a7e417f13e4d",
11
  "environment": {
12
+ "python": "3.11.18",
13
+ "transformers": "5.21.0.dev0",
14
+ "torch": "2.14.0+cu130",
15
+ "torchvision": "0.29.1+cu130",
16
+ "torchaudio": "2.11.0+cu130",
17
+ "tokenizers": "0.23.3",
18
+ "numpy": "2.4.6",
19
+ "pillow": "13.0.0",
20
+ "accelerate": "1.16.0.dev0",
21
  "safetensors": "0.8.0",
22
+ "huggingface-hub": "2.3.0",
23
+ "sentencepiece": "0.2.2",
24
+ "protobuf": "7.36.2",
25
  "librosa": "0.11.0",
26
+ "soundfile": "0.14.0",
27
+ "scipy": "1.17.1"
28
+ },
29
+ "seconds": 7.34,
30
+ "environment_differs": {
31
+ "torch": "2.14.0+cu130 (pinned 2.11.0)",
32
+ "torchvision": "0.29.1+cu130 (pinned 0.26.0)",
33
+ "tokenizers": "0.23.3 (pinned 0.23.1)",
34
+ "numpy": "2.4.6 (pinned 1.26.4)",
35
+ "pillow": "13.0.0 (pinned 12.2.0)",
36
+ "accelerate": "1.16.0.dev0 (pinned 1.13.0)",
37
+ "huggingface-hub": "2.3.0 (pinned 1.32.0)",
38
+ "sentencepiece": "0.2.2 (pinned 0.2.1)",
39
+ "protobuf": "7.36.2 (pinned 6.33.6)",
40
+ "soundfile": "0.14.0 (pinned 0.13.1)",
41
+ "scipy": "1.17.1 (pinned 1.15.3)"
42
  },
 
43
  "platform": "cuda"
44
  },
45
  "outputs": {
integrations/transformers/expectations/cuda/detr-mic.eager.json CHANGED
@@ -1,32 +1,45 @@
1
  {
2
  "recorded_with": {
3
- "contract": "hf-internal-testing/mic-detr-integration",
4
- "contract_name": "detr",
5
- "contract_revision": "unpublished",
6
  "framework": "transformers",
7
  "variant": "eager",
8
  "target": "detr-mic",
9
  "kind": "fixture",
10
- "revision": "unpublished",
11
  "environment": {
12
- "python": "3.10.12",
13
- "transformers": "5.19.0.dev0",
14
- "torch": "2.11.0+cu128",
15
- "torchvision": "0.26.0+cu128",
16
- "torchaudio": "2.11.0+cu128",
17
- "tokenizers": "0.23.1",
18
- "numpy": "1.26.4",
19
- "pillow": "12.2.0",
20
- "accelerate": "1.13.0",
21
  "safetensors": "0.8.0",
22
- "huggingface-hub": "1.32.0",
23
- "sentencepiece": "0.2.1",
24
- "protobuf": "6.33.6",
25
  "librosa": "0.11.0",
26
- "soundfile": "0.13.1",
27
- "scipy": "1.15.3"
 
 
 
 
 
 
 
 
 
 
 
 
 
 
28
  },
29
- "seconds": 5.01,
30
  "platform": "cuda"
31
  },
32
  "outputs": {
@@ -37,49 +50,49 @@
37
  },
38
  "top_class_logits": [
39
  [
40
- -0.13664180040359497,
41
- 0.03382372856140137,
42
- -0.2539277672767639,
43
- 0.1532427817583084,
44
- 0.042901769280433655
45
  ],
46
  [
47
- -0.1366385817527771,
48
- 0.03382717818021774,
49
- -0.25393906235694885,
50
- 0.15323668718338013,
51
- 0.042901962995529175
52
  ],
53
  [
54
- -0.13664469122886658,
55
- 0.03383926674723625,
56
- -0.25393587350845337,
57
- 0.1532299816608429,
58
- 0.042911700904369354
59
  ]
60
  ],
61
  "pred_boxes": [
62
  [
63
- 0.49977418780326843,
64
- 0.5002057552337646,
65
- 0.49946823716163635,
66
- 0.4998737871646881
67
  ],
68
  [
69
- 0.49977418780326843,
70
- 0.5002058148384094,
71
- 0.4994681775569916,
72
- 0.4998737871646881
73
  ],
74
  [
75
- 0.49977418780326843,
76
- 0.5002058148384094,
77
- 0.49946823716163635,
78
- 0.4998737871646881
79
  ]
80
  ],
81
  "training": {
82
- "loss": 8.983428955078125,
83
  "gradients": {
84
  "model.backbone.model.conv1.weight": true,
85
  "model.backbone.model.layer1.0.downsample.0.weight": true
 
1
  {
2
  "recorded_with": {
3
+ "contract": "huggingface-mic/detr-mic",
4
+ "contract_name": "mic-detr",
5
+ "contract_revision": "91ff508188db759f2ee39ec64016a7e417f13e4d",
6
  "framework": "transformers",
7
  "variant": "eager",
8
  "target": "detr-mic",
9
  "kind": "fixture",
10
+ "revision": "91ff508188db759f2ee39ec64016a7e417f13e4d",
11
  "environment": {
12
+ "python": "3.11.18",
13
+ "transformers": "5.21.0.dev0",
14
+ "torch": "2.14.0+cu130",
15
+ "torchvision": "0.29.1+cu130",
16
+ "torchaudio": "2.11.0+cu130",
17
+ "tokenizers": "0.23.3",
18
+ "numpy": "2.4.6",
19
+ "pillow": "13.0.0",
20
+ "accelerate": "1.16.0.dev0",
21
  "safetensors": "0.8.0",
22
+ "huggingface-hub": "2.3.0",
23
+ "sentencepiece": "0.2.2",
24
+ "protobuf": "7.36.2",
25
  "librosa": "0.11.0",
26
+ "soundfile": "0.14.0",
27
+ "scipy": "1.17.1"
28
+ },
29
+ "seconds": 7.09,
30
+ "environment_differs": {
31
+ "torch": "2.14.0+cu130 (pinned 2.11.0)",
32
+ "torchvision": "0.29.1+cu130 (pinned 0.26.0)",
33
+ "tokenizers": "0.23.3 (pinned 0.23.1)",
34
+ "numpy": "2.4.6 (pinned 1.26.4)",
35
+ "pillow": "13.0.0 (pinned 12.2.0)",
36
+ "accelerate": "1.16.0.dev0 (pinned 1.13.0)",
37
+ "huggingface-hub": "2.3.0 (pinned 1.32.0)",
38
+ "sentencepiece": "0.2.2 (pinned 0.2.1)",
39
+ "protobuf": "7.36.2 (pinned 6.33.6)",
40
+ "soundfile": "0.14.0 (pinned 0.13.1)",
41
+ "scipy": "1.17.1 (pinned 1.15.3)"
42
  },
 
43
  "platform": "cuda"
44
  },
45
  "outputs": {
 
50
  },
51
  "top_class_logits": [
52
  [
53
+ -0.1366158425807953,
54
+ 0.0338565856218338,
55
+ -0.2539196014404297,
56
+ 0.15320105850696564,
57
+ 0.04292171448469162
58
  ],
59
  [
60
+ -0.13661256432533264,
61
+ 0.03386005759239197,
62
+ -0.2539310157299042,
63
+ 0.15319499373435974,
64
+ 0.04292191192507744
65
  ],
66
  [
67
+ -0.1366187334060669,
68
+ 0.033872127532958984,
69
+ -0.2539277672767639,
70
+ 0.15318834781646729,
71
+ 0.04293164238333702
72
  ]
73
  ],
74
  "pred_boxes": [
75
  [
76
+ 0.4997738301753998,
77
+ 0.5002056956291199,
78
+ 0.4994681775569916,
79
+ 0.49987390637397766
80
  ],
81
  [
82
+ 0.4997738301753998,
83
+ 0.5002057552337646,
84
+ 0.499468058347702,
85
+ 0.49987390637397766
86
  ],
87
  [
88
+ 0.4997738301753998,
89
+ 0.5002057552337646,
90
+ 0.4994681775569916,
91
+ 0.49987390637397766
92
  ]
93
  ],
94
  "training": {
95
+ "loss": 8.983436584472656,
96
  "gradients": {
97
  "model.backbone.model.conv1.weight": true,
98
  "model.backbone.model.layer1.0.downsample.0.weight": true
integrations/transformers/expectations/cuda/detr-mic.json CHANGED
@@ -1,32 +1,45 @@
1
  {
2
  "recorded_with": {
3
- "contract": "hf-internal-testing/mic-detr-integration",
4
- "contract_name": "detr",
5
- "contract_revision": "unpublished",
6
  "framework": "transformers",
7
  "variant": "default",
8
  "target": "detr-mic",
9
  "kind": "fixture",
10
- "revision": "unpublished",
11
  "environment": {
12
- "python": "3.10.12",
13
- "transformers": "5.19.0.dev0",
14
- "torch": "2.11.0+cu128",
15
- "torchvision": "0.26.0+cu128",
16
- "torchaudio": "2.11.0+cu128",
17
- "tokenizers": "0.23.1",
18
- "numpy": "1.26.4",
19
- "pillow": "12.2.0",
20
- "accelerate": "1.13.0",
21
  "safetensors": "0.8.0",
22
- "huggingface-hub": "1.32.0",
23
- "sentencepiece": "0.2.1",
24
- "protobuf": "6.33.6",
25
  "librosa": "0.11.0",
26
- "soundfile": "0.13.1",
27
- "scipy": "1.15.3"
 
 
 
 
 
 
 
 
 
 
 
 
 
 
28
  },
29
- "seconds": 5.07,
30
  "platform": "cuda"
31
  },
32
  "outputs": {
@@ -37,49 +50,49 @@
37
  },
38
  "top_class_logits": [
39
  [
40
- -0.13664181530475616,
41
- 0.03382376581430435,
42
- -0.2539277672767639,
43
- 0.1532427966594696,
44
- 0.04290176182985306
45
  ],
46
  [
47
- -0.1366385817527771,
48
- 0.033827196806669235,
49
- -0.25393903255462646,
50
- 0.15323670208454132,
51
- 0.04290197417140007
52
  ],
53
  [
54
- -0.13664475083351135,
55
- 0.033839281648397446,
56
- -0.25393590331077576,
57
- 0.15322993695735931,
58
- 0.04291171208024025
59
  ]
60
  ],
61
  "pred_boxes": [
62
  [
63
- 0.49977418780326843,
64
- 0.5002057552337646,
65
- 0.49946823716163635,
66
- 0.4998737871646881
67
  ],
68
  [
69
- 0.49977418780326843,
70
- 0.5002058148384094,
71
- 0.4994681775569916,
72
- 0.4998737871646881
73
  ],
74
  [
75
- 0.49977418780326843,
76
- 0.5002058148384094,
77
- 0.49946823716163635,
78
- 0.4998737871646881
79
  ]
80
  ],
81
  "training": {
82
- "loss": 8.983428955078125,
83
  "gradients": {
84
  "model.backbone.model.conv1.weight": true,
85
  "model.backbone.model.layer1.0.downsample.0.weight": true
 
1
  {
2
  "recorded_with": {
3
+ "contract": "huggingface-mic/detr-mic",
4
+ "contract_name": "mic-detr",
5
+ "contract_revision": "91ff508188db759f2ee39ec64016a7e417f13e4d",
6
  "framework": "transformers",
7
  "variant": "default",
8
  "target": "detr-mic",
9
  "kind": "fixture",
10
+ "revision": "91ff508188db759f2ee39ec64016a7e417f13e4d",
11
  "environment": {
12
+ "python": "3.11.18",
13
+ "transformers": "5.21.0.dev0",
14
+ "torch": "2.14.0+cu130",
15
+ "torchvision": "0.29.1+cu130",
16
+ "torchaudio": "2.11.0+cu130",
17
+ "tokenizers": "0.23.3",
18
+ "numpy": "2.4.6",
19
+ "pillow": "13.0.0",
20
+ "accelerate": "1.16.0.dev0",
21
  "safetensors": "0.8.0",
22
+ "huggingface-hub": "2.3.0",
23
+ "sentencepiece": "0.2.2",
24
+ "protobuf": "7.36.2",
25
  "librosa": "0.11.0",
26
+ "soundfile": "0.14.0",
27
+ "scipy": "1.17.1"
28
+ },
29
+ "seconds": 7.14,
30
+ "environment_differs": {
31
+ "torch": "2.14.0+cu130 (pinned 2.11.0)",
32
+ "torchvision": "0.29.1+cu130 (pinned 0.26.0)",
33
+ "tokenizers": "0.23.3 (pinned 0.23.1)",
34
+ "numpy": "2.4.6 (pinned 1.26.4)",
35
+ "pillow": "13.0.0 (pinned 12.2.0)",
36
+ "accelerate": "1.16.0.dev0 (pinned 1.13.0)",
37
+ "huggingface-hub": "2.3.0 (pinned 1.32.0)",
38
+ "sentencepiece": "0.2.2 (pinned 0.2.1)",
39
+ "protobuf": "7.36.2 (pinned 6.33.6)",
40
+ "soundfile": "0.14.0 (pinned 0.13.1)",
41
+ "scipy": "1.17.1 (pinned 1.15.3)"
42
  },
 
43
  "platform": "cuda"
44
  },
45
  "outputs": {
 
50
  },
51
  "top_class_logits": [
52
  [
53
+ -0.1366157978773117,
54
+ 0.0338565856218338,
55
+ -0.2539196014404297,
56
+ 0.1532011330127716,
57
+ 0.04292171448469162
58
  ],
59
  [
60
+ -0.13661256432533264,
61
+ 0.033860038965940475,
62
+ -0.2539309859275818,
63
+ 0.15319503843784332,
64
+ 0.042921893298625946
65
  ],
66
  [
67
+ -0.13661877810955048,
68
+ 0.03387211635708809,
69
+ -0.2539277672767639,
70
+ 0.15318837761878967,
71
+ 0.042931683361530304
72
  ]
73
  ],
74
  "pred_boxes": [
75
  [
76
+ 0.4997738301753998,
77
+ 0.5002056956291199,
78
+ 0.4994681775569916,
79
+ 0.49987390637397766
80
  ],
81
  [
82
+ 0.4997738301753998,
83
+ 0.5002057552337646,
84
+ 0.499468058347702,
85
+ 0.49987390637397766
86
  ],
87
  [
88
+ 0.4997738301753998,
89
+ 0.5002057552337646,
90
+ 0.4994681775569916,
91
+ 0.49987390637397766
92
  ]
93
  ],
94
  "training": {
95
+ "loss": 8.983436584472656,
96
  "gradients": {
97
  "model.backbone.model.conv1.weight": true,
98
  "model.backbone.model.layer1.0.downsample.0.weight": true