File size: 3,805 Bytes
e4ab6c6 | 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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 | {
"compute": {
"environment": "Linux x86_64 CPU with mlx[cpu]; no paid cloud Mac used.",
"estimated_incremental_cost_usd": 0.0
},
"end_to_end_extraction": {
"chunk_lengths": [
50
],
"input_html": "<html><body><article><h1>Apple MLX conversion</h1><p>This article explains how to convert a EuroBERT content extraction model to MLX format.</p></article></body></html>",
"num_blocks": 2,
"variants": {
"4bit": {
"html": "<html><body><article><h1 _item_id=\"1\">Apple MLX conversion</h1><p _item_id=\"2\">This article explains how to convert a EuroBERT content extraction model to MLX format.</p></article></body></html>",
"labels": {
"1": "main",
"2": "main"
},
"latency_ms": 77852.023473999,
"markdown": "<html><body><article><h1 _item_id=\"1\">Apple MLX conversion</h1><p _item_id=\"2\">This article explains how to convert a EuroBERT content extraction model to MLX format.</p></article></body></html>",
"non_empty": true,
"predictions": [
1,
1
]
},
"8bit": {
"html": "<html><body><article><h1 _item_id=\"1\">Apple MLX conversion</h1><p _item_id=\"2\">This article explains how to convert a EuroBERT content extraction model to MLX format.</p></article></body></html>",
"labels": {
"1": "main",
"2": "main"
},
"latency_ms": 78333.6143399938,
"markdown": "<html><body><article><h1 _item_id=\"1\">Apple MLX conversion</h1><p _item_id=\"2\">This article explains how to convert a EuroBERT content extraction model to MLX format.</p></article></body></html>",
"non_empty": true,
"predictions": [
1,
1
]
},
"bf16": {
"html": "<html><body><article><h1 _item_id=\"1\">Apple MLX conversion</h1><p _item_id=\"2\">This article explains how to convert a EuroBERT content extraction model to MLX format.</p></article></body></html>",
"labels": {
"1": "main",
"2": "main"
},
"latency_ms": 5469.0545329940505,
"markdown": "<html><body><article><h1 _item_id=\"1\">Apple MLX conversion</h1><p _item_id=\"2\">This article explains how to convert a EuroBERT content extraction model to MLX format.</p></article></body></html>",
"non_empty": true,
"predictions": [
1,
1
]
}
}
},
"load_checks": {
"4bit": {
"loaded": true,
"logits_dtype": "mlx.core.bfloat16",
"logits_shape": [
1,
3,
2
]
},
"8bit": {
"loaded": true,
"logits_dtype": "mlx.core.bfloat16",
"logits_shape": [
1,
3,
2
]
},
"bf16": {
"loaded": true,
"logits_dtype": "mlx.core.bfloat16",
"logits_shape": [
1,
3,
2
]
}
},
"model_dir": "hf_out",
"numerical_accuracy": {
"test_inputs": [
"A",
"B",
"C"
],
"token_shape": [
3,
2
],
"torch_reference": {
"attention": "eager",
"dtype": "float32",
"latency_ms": 50.09009799687192
},
"variants": {
"4bit": {
"latency_ms": 9323.177823971491,
"max_abs_diff": 2.2551989555358887,
"mean_abs_diff": 1.1897996664047241
},
"8bit": {
"latency_ms": 9380.578643002082,
"max_abs_diff": 1.2797489166259766,
"mean_abs_diff": 0.542361319065094
},
"bf16": {
"latency_ms": 716.8781639775261,
"max_abs_diff": 0.04523277282714844,
"mean_abs_diff": 0.019181733950972557
}
}
},
"source_model": "feyninc/pulpie-orange-small",
"variants": [
"bf16",
"8bit",
"4bit"
]
}
|