Instructions to use spybyscript/gliner25-decide-litert with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- LiteRT
How to use spybyscript/gliner25-decide-litert with LiteRT:
# No code snippets available yet for this library. # To use this model, check the repository files and the library's documentation. # Want to help? PRs adding snippets are welcome at: # https://github.com/huggingface/huggingface.js
- GLiNER2
How to use spybyscript/gliner25-decide-litert with GLiNER2:
from gliner2 import AutoExtractor extractor = AutoExtractor.from_pretrained("spybyscript/gliner25-decide-litert") # Extract entities text = "Apple CEO Tim Cook announced iPhone 15 in Cupertino yesterday." result = extractor.extract_entities(text, ["company", "person", "product", "location"]) print(result) - Notebooks
- Google Colab
- Kaggle
Download validation.json from spybyscript/gliner25-decide-litert: direct link, hf CLI and curl.
- Browser
- Download file 27.6 kB
-
https://huggingface.co/spybyscript/gliner25-decide-litert/resolve/main/validation.json
- Command line
-
hf download hf://spybyscript/gliner25-decide-litert/validation.json
-
curl -L -o validation.json https://huggingface.co/spybyscript/gliner25-decide-litert/resolve/main/validation.json
27.6 kB
| { | |
| "model": "gliner25-decide-litert", | |
| "source": "fastino/GLiNER2.5-Decide@bbe10ff77ebb238777c17d3a8ac9260e30929057", | |
| "date": "2026-10-28", | |
| "versions": { | |
| "litert-torch": "0.9.4", | |
| "ai-edge-litert": "2.2.0", | |
| "ai-edge-quantizer": "0.9.0", | |
| "litert-converter": "0.4.0", | |
| "torch": "2.13.0", | |
| "transformers": "4.57.6", | |
| "gliner2": "2.0.0" | |
| }, | |
| "machine": "arm64 Mac (Apple M5 Max, macOS 27.0.1)", | |
| "buckets": [ | |
| 128, | |
| 160, | |
| 256, | |
| 512 | |
| ], | |
| "threads": 4, | |
| "corpus": { | |
| "calls": 616, | |
| "quick": false, | |
| "source": "gliner2 2.0.0 library calls recorded on a synthetic English corpus of short conversational statements", | |
| "byBucket": { | |
| "128": 8, | |
| "160": 321, | |
| "512": 287 | |
| } | |
| }, | |
| "gate": { | |
| "agreement": 0.95, | |
| "drift": 0.05 | |
| }, | |
| "variants": { | |
| "wfp16": { | |
| "about": "FULLY_CONNECTED weights stored float16 behind DEQUANTIZE (ai-edge-quantizer FLOAT_CASTING); FP32 activations, relative-position tables, norms and biases; finite -1e4 mask", | |
| "file": "gliner25_decide_wfp16.tflite", | |
| "bytes": 1026117712, | |
| "buildSeconds": 5.6, | |
| "tableSharing": null, | |
| "tablesFp16": null, | |
| "audit": { | |
| "bytes": 1026117712, | |
| "signatures": [ | |
| { | |
| "key": "s128", | |
| "inputs": [ | |
| { | |
| "name": "attention_mask", | |
| "dtype": "float32", | |
| "shape": [ | |
| 1, | |
| 128 | |
| ] | |
| }, | |
| { | |
| "name": "inputs_embeds", | |
| "dtype": "float32", | |
| "shape": [ | |
| 1, | |
| 128, | |
| 1024 | |
| ] | |
| } | |
| ], | |
| "outputs": [ | |
| { | |
| "name": "logits", | |
| "dtype": "float32", | |
| "shape": [ | |
| 1, | |
| 128 | |
| ] | |
| } | |
| ] | |
| }, | |
| { | |
| "key": "s160", | |
| "inputs": [ | |
| { | |
| "name": "attention_mask", | |
| "dtype": "float32", | |
| "shape": [ | |
| 1, | |
| 160 | |
| ] | |
| }, | |
| { | |
| "name": "inputs_embeds", | |
| "dtype": "float32", | |
| "shape": [ | |
| 1, | |
| 160, | |
| 1024 | |
| ] | |
| } | |
| ], | |
| "outputs": [ | |
| { | |
| "name": "logits", | |
| "dtype": "float32", | |
| "shape": [ | |
| 1, | |
| 160 | |
| ] | |
| } | |
| ] | |
| }, | |
| { | |
| "key": "s256", | |
| "inputs": [ | |
| { | |
| "name": "attention_mask", | |
| "dtype": "float32", | |
| "shape": [ | |
| 1, | |
| 256 | |
| ] | |
| }, | |
| { | |
| "name": "inputs_embeds", | |
| "dtype": "float32", | |
| "shape": [ | |
| 1, | |
| 256, | |
| 1024 | |
| ] | |
| } | |
| ], | |
| "outputs": [ | |
| { | |
| "name": "logits", | |
| "dtype": "float32", | |
| "shape": [ | |
| 1, | |
| 256 | |
| ] | |
| } | |
| ] | |
| }, | |
| { | |
| "key": "s512", | |
| "inputs": [ | |
| { | |
| "name": "attention_mask", | |
| "dtype": "float32", | |
| "shape": [ | |
| 1, | |
| 512 | |
| ] | |
| }, | |
| { | |
| "name": "inputs_embeds", | |
| "dtype": "float32", | |
| "shape": [ | |
| 1, | |
| 512, | |
| 1024 | |
| ] | |
| } | |
| ], | |
| "outputs": [ | |
| { | |
| "name": "logits", | |
| "dtype": "float32", | |
| "shape": [ | |
| 1, | |
| 512 | |
| ] | |
| } | |
| ] | |
| } | |
| ], | |
| "operators": { | |
| "s128": 1749, | |
| "s160": 1749, | |
| "s256": 1749, | |
| "s512": 1749 | |
| }, | |
| "opHistogram": { | |
| "s128": { | |
| "ADD": 266, | |
| "BATCH_MATMUL": 96, | |
| "DEQUANTIZE": 146, | |
| "FULLY_CONNECTED": 146, | |
| "GELU": 24, | |
| "MEAN": 98, | |
| "MUL": 220, | |
| "RESHAPE": 341, | |
| "RSQRT": 49, | |
| "SLICE": 96, | |
| "SOFTMAX": 24, | |
| "SQUARED_DIFFERENCE": 49, | |
| "SUB": 50, | |
| "TRANSPOSE": 144 | |
| }, | |
| "s160": { | |
| "ADD": 266, | |
| "BATCH_MATMUL": 96, | |
| "DEQUANTIZE": 146, | |
| "FULLY_CONNECTED": 146, | |
| "GELU": 24, | |
| "MEAN": 98, | |
| "MUL": 220, | |
| "RESHAPE": 341, | |
| "RSQRT": 49, | |
| "SLICE": 96, | |
| "SOFTMAX": 24, | |
| "SQUARED_DIFFERENCE": 49, | |
| "SUB": 50, | |
| "TRANSPOSE": 144 | |
| }, | |
| "s256": { | |
| "ADD": 266, | |
| "BATCH_MATMUL": 96, | |
| "DEQUANTIZE": 146, | |
| "FULLY_CONNECTED": 146, | |
| "GELU": 24, | |
| "MEAN": 98, | |
| "MUL": 220, | |
| "RESHAPE": 341, | |
| "RSQRT": 49, | |
| "SLICE": 96, | |
| "SOFTMAX": 24, | |
| "SQUARED_DIFFERENCE": 49, | |
| "SUB": 50, | |
| "TRANSPOSE": 144 | |
| }, | |
| "s512": { | |
| "ADD": 266, | |
| "BATCH_MATMUL": 96, | |
| "DEQUANTIZE": 146, | |
| "FULLY_CONNECTED": 146, | |
| "GELU": 24, | |
| "MEAN": 98, | |
| "MUL": 220, | |
| "RESHAPE": 341, | |
| "RSQRT": 49, | |
| "SLICE": 96, | |
| "SOFTMAX": 24, | |
| "SQUARED_DIFFERENCE": 49, | |
| "SUB": 50, | |
| "TRANSPOSE": 144 | |
| } | |
| }, | |
| "tensorDtypes": { | |
| "FLOAT32": 8184, | |
| "FLOAT16": 584, | |
| "INT32": 60 | |
| }, | |
| "int64Tensors": 0, | |
| "constantBytesBySignatures": { | |
| "s128+s160+s256+s512": 609464416, | |
| "s512": 201326716, | |
| "s256": 100663420, | |
| "s160": 62914684, | |
| "s128": 50331772 | |
| }, | |
| "activationsWithLeadingDimNot1": [ | |
| { | |
| "signature": "s128", | |
| "shape": [ | |
| 128, | |
| 1024 | |
| ] | |
| }, | |
| { | |
| "signature": "s128", | |
| "shape": [ | |
| 128, | |
| 1024 | |
| ] | |
| }, | |
| { | |
| "signature": "s128", | |
| "shape": [ | |
| 128, | |
| 1024 | |
| ] | |
| }, | |
| { | |
| "signature": "s128", | |
| "shape": [ | |
| 128, | |
| 1024 | |
| ] | |
| }, | |
| { | |
| "signature": "s128", | |
| "shape": [ | |
| 128, | |
| 1024 | |
| ] | |
| }, | |
| { | |
| "signature": "s128", | |
| "shape": [ | |
| 128, | |
| 1024 | |
| ] | |
| }, | |
| { | |
| "signature": "s128", | |
| "shape": [ | |
| 128, | |
| 1024 | |
| ] | |
| }, | |
| { | |
| "signature": "s128", | |
| "shape": [ | |
| 128, | |
| 1024 | |
| ] | |
| } | |
| ] | |
| }, | |
| "parity": { | |
| "cpu": { | |
| "calls": 616, | |
| "decisions": 2338, | |
| "agree": 2338, | |
| "agreement": 1.0, | |
| "driftMax": 0.00032, | |
| "driftP99": 0.00017, | |
| "driftP50": 1.7e-05, | |
| "over005": 0, | |
| "msByBucket": { | |
| "128": { | |
| "calls": 8, | |
| "p50": 58.3, | |
| "p95": 64.8 | |
| }, | |
| "160": { | |
| "calls": 321, | |
| "p50": 68.1, | |
| "p95": 70.5 | |
| }, | |
| "512": { | |
| "calls": 287, | |
| "p50": 248.6, | |
| "p95": 255.8 | |
| } | |
| }, | |
| "loadSeconds": 0.53, | |
| "fullyAccelerated": true, | |
| "gate": "pass" | |
| }, | |
| "gpu_fp32": { | |
| "calls": 616, | |
| "decisions": 2338, | |
| "agree": 2338, | |
| "agreement": 1.0, | |
| "driftMax": 0.00032, | |
| "driftP99": 0.00017, | |
| "driftP50": 1.7e-05, | |
| "over005": 0, | |
| "msByBucket": { | |
| "128": { | |
| "calls": 8, | |
| "p50": 21.0, | |
| "p95": 21.9 | |
| }, | |
| "160": { | |
| "calls": 321, | |
| "p50": 23.8, | |
| "p95": 25.4 | |
| }, | |
| "512": { | |
| "calls": 287, | |
| "p50": 76.1, | |
| "p95": 79.5 | |
| } | |
| }, | |
| "loadSeconds": 1.8, | |
| "fullyAccelerated": true, | |
| "gate": "pass" | |
| }, | |
| "gpu_fp16": { | |
| "calls": 616, | |
| "decisions": 2338, | |
| "agree": 2226, | |
| "agreement": 0.9521, | |
| "driftMax": 0.30555, | |
| "driftP99": 0.12811, | |
| "driftP50": 0.012985, | |
| "over005": 241, | |
| "msByBucket": { | |
| "128": { | |
| "calls": 8, | |
| "p50": 16.3, | |
| "p95": 16.3 | |
| }, | |
| "160": { | |
| "calls": 321, | |
| "p50": 20.5, | |
| "p95": 20.8 | |
| }, | |
| "512": { | |
| "calls": 287, | |
| "p50": 49.6, | |
| "p95": 49.8 | |
| } | |
| }, | |
| "loadSeconds": 1.73, | |
| "fullyAccelerated": true, | |
| "gate": [ | |
| "probability drift 0.30555" | |
| ] | |
| } | |
| }, | |
| "logitDifferenceFromTorch": { | |
| "cpu": { | |
| "160": 0.000954, | |
| "512": 0.0015 | |
| }, | |
| "gpu_fp32": { | |
| "160": 0.000952, | |
| "512": 0.00151 | |
| }, | |
| "gpu_fp16": { | |
| "160": 1.24, | |
| "512": 0.556 | |
| } | |
| }, | |
| "memory": { | |
| "cpuNoCache": { | |
| "loadSeconds": 0.62, | |
| "baseline": { | |
| "rssMb": 194.2, | |
| "footprintMb": 149.3 | |
| }, | |
| "afterLoad": { | |
| "rssMb": 6374.3, | |
| "footprintMb": 5348.1 | |
| }, | |
| "afterEachSignature": { | |
| "s128": { | |
| "rssMb": 6385.7, | |
| "footprintMb": 5359.0 | |
| }, | |
| "s160": { | |
| "rssMb": 6403.3, | |
| "footprintMb": 5376.6 | |
| }, | |
| "s256": { | |
| "rssMb": 6441.6, | |
| "footprintMb": 5414.9 | |
| }, | |
| "s512": { | |
| "rssMb": 6583.8, | |
| "footprintMb": 5557.2 | |
| } | |
| }, | |
| "peakRssMb": 6583.8 | |
| }, | |
| "cpuCacheFirst": { | |
| "loadSeconds": 0.33, | |
| "baseline": { | |
| "rssMb": 194.2, | |
| "footprintMb": 149.3 | |
| }, | |
| "afterLoad": { | |
| "rssMb": 1708.6, | |
| "footprintMb": 679.0 | |
| }, | |
| "afterEachSignature": { | |
| "s128": { | |
| "rssMb": 2871.1, | |
| "footprintMb": 680.4 | |
| }, | |
| "s160": { | |
| "rssMb": 2888.9, | |
| "footprintMb": 698.3 | |
| }, | |
| "s256": { | |
| "rssMb": 2927.2, | |
| "footprintMb": 736.6 | |
| }, | |
| "s512": { | |
| "rssMb": 3069.4, | |
| "footprintMb": 878.8 | |
| } | |
| }, | |
| "peakRssMb": 3069.4 | |
| }, | |
| "cpuCacheWarm": { | |
| "loadSeconds": 0.22, | |
| "baseline": { | |
| "rssMb": 194.5, | |
| "footprintMb": 149.5 | |
| }, | |
| "afterLoad": { | |
| "rssMb": 1691.0, | |
| "footprintMb": 661.5 | |
| }, | |
| "afterEachSignature": { | |
| "s128": { | |
| "rssMb": 2853.5, | |
| "footprintMb": 662.9 | |
| }, | |
| "s160": { | |
| "rssMb": 2871.3, | |
| "footprintMb": 680.7 | |
| }, | |
| "s256": { | |
| "rssMb": 2909.6, | |
| "footprintMb": 719.0 | |
| }, | |
| "s512": { | |
| "rssMb": 3051.8, | |
| "footprintMb": 861.4 | |
| } | |
| }, | |
| "peakRssMb": 3051.8 | |
| } | |
| } | |
| }, | |
| "fp16hw": { | |
| "about": "wfp16's storage (float16 FULLY_CONNECTED weights behind DEQUANTIZE) over the FP16-hardware graph: LayerNorm power-of-two pre-scale (x*2^-k, eps*4^-k, k from the corpus' FP32 ranges), first feed-forward FULLY_CONNECTED split in two halves, finite -1e4 mask", | |
| "file": "gliner25_decide_fp16hw.tflite", | |
| "bytes": 1026073648, | |
| "buildSeconds": 5.7, | |
| "tableSharing": null, | |
| "tablesFp16": null, | |
| "audit": { | |
| "bytes": 1026073648, | |
| "signatures": [ | |
| { | |
| "key": "s128", | |
| "inputs": [ | |
| { | |
| "name": "attention_mask", | |
| "dtype": "float32", | |
| "shape": [ | |
| 1, | |
| 128 | |
| ] | |
| }, | |
| { | |
| "name": "inputs_embeds", | |
| "dtype": "float32", | |
| "shape": [ | |
| 1, | |
| 128, | |
| 1024 | |
| ] | |
| } | |
| ], | |
| "outputs": [ | |
| { | |
| "name": "logits", | |
| "dtype": "float32", | |
| "shape": [ | |
| 1, | |
| 128 | |
| ] | |
| } | |
| ] | |
| }, | |
| { | |
| "key": "s160", | |
| "inputs": [ | |
| { | |
| "name": "attention_mask", | |
| "dtype": "float32", | |
| "shape": [ | |
| 1, | |
| 160 | |
| ] | |
| }, | |
| { | |
| "name": "inputs_embeds", | |
| "dtype": "float32", | |
| "shape": [ | |
| 1, | |
| 160, | |
| 1024 | |
| ] | |
| } | |
| ], | |
| "outputs": [ | |
| { | |
| "name": "logits", | |
| "dtype": "float32", | |
| "shape": [ | |
| 1, | |
| 160 | |
| ] | |
| } | |
| ] | |
| }, | |
| { | |
| "key": "s256", | |
| "inputs": [ | |
| { | |
| "name": "attention_mask", | |
| "dtype": "float32", | |
| "shape": [ | |
| 1, | |
| 256 | |
| ] | |
| }, | |
| { | |
| "name": "inputs_embeds", | |
| "dtype": "float32", | |
| "shape": [ | |
| 1, | |
| 256, | |
| 1024 | |
| ] | |
| } | |
| ], | |
| "outputs": [ | |
| { | |
| "name": "logits", | |
| "dtype": "float32", | |
| "shape": [ | |
| 1, | |
| 256 | |
| ] | |
| } | |
| ] | |
| }, | |
| { | |
| "key": "s512", | |
| "inputs": [ | |
| { | |
| "name": "attention_mask", | |
| "dtype": "float32", | |
| "shape": [ | |
| 1, | |
| 512 | |
| ] | |
| }, | |
| { | |
| "name": "inputs_embeds", | |
| "dtype": "float32", | |
| "shape": [ | |
| 1, | |
| 512, | |
| 1024 | |
| ] | |
| } | |
| ], | |
| "outputs": [ | |
| { | |
| "name": "logits", | |
| "dtype": "float32", | |
| "shape": [ | |
| 1, | |
| 512 | |
| ] | |
| } | |
| ] | |
| } | |
| ], | |
| "operators": { | |
| "s128": 1846, | |
| "s160": 1846, | |
| "s256": 1846, | |
| "s512": 1846 | |
| }, | |
| "opHistogram": { | |
| "s128": { | |
| "ADD": 266, | |
| "BATCH_MATMUL": 96, | |
| "CONCATENATION": 24, | |
| "DEQUANTIZE": 170, | |
| "FULLY_CONNECTED": 170, | |
| "GELU": 24, | |
| "MEAN": 98, | |
| "MUL": 245, | |
| "RESHAPE": 341, | |
| "RSQRT": 49, | |
| "SLICE": 96, | |
| "SOFTMAX": 24, | |
| "SQUARED_DIFFERENCE": 49, | |
| "SUB": 50, | |
| "TRANSPOSE": 144 | |
| }, | |
| "s160": { | |
| "ADD": 266, | |
| "BATCH_MATMUL": 96, | |
| "CONCATENATION": 24, | |
| "DEQUANTIZE": 170, | |
| "FULLY_CONNECTED": 170, | |
| "GELU": 24, | |
| "MEAN": 98, | |
| "MUL": 245, | |
| "RESHAPE": 341, | |
| "RSQRT": 49, | |
| "SLICE": 96, | |
| "SOFTMAX": 24, | |
| "SQUARED_DIFFERENCE": 49, | |
| "SUB": 50, | |
| "TRANSPOSE": 144 | |
| }, | |
| "s256": { | |
| "ADD": 266, | |
| "BATCH_MATMUL": 96, | |
| "CONCATENATION": 24, | |
| "DEQUANTIZE": 170, | |
| "FULLY_CONNECTED": 170, | |
| "GELU": 24, | |
| "MEAN": 98, | |
| "MUL": 245, | |
| "RESHAPE": 341, | |
| "RSQRT": 49, | |
| "SLICE": 96, | |
| "SOFTMAX": 24, | |
| "SQUARED_DIFFERENCE": 49, | |
| "SUB": 50, | |
| "TRANSPOSE": 144 | |
| }, | |
| "s512": { | |
| "ADD": 266, | |
| "BATCH_MATMUL": 96, | |
| "CONCATENATION": 24, | |
| "DEQUANTIZE": 170, | |
| "FULLY_CONNECTED": 170, | |
| "GELU": 24, | |
| "MEAN": 98, | |
| "MUL": 245, | |
| "RESHAPE": 341, | |
| "RSQRT": 49, | |
| "SLICE": 96, | |
| "SOFTMAX": 24, | |
| "SQUARED_DIFFERENCE": 49, | |
| "SUB": 50, | |
| "TRANSPOSE": 144 | |
| } | |
| }, | |
| "tensorDtypes": { | |
| "FLOAT32": 8716, | |
| "FLOAT16": 680, | |
| "INT32": 60 | |
| }, | |
| "int64Tensors": 0, | |
| "constantBytesBySignatures": { | |
| "s128+s160+s256+s512": 609464464, | |
| "s512": 201326716, | |
| "s256": 100663420, | |
| "s160": 62914684, | |
| "s128": 50331772 | |
| }, | |
| "activationsWithLeadingDimNot1": [ | |
| { | |
| "signature": "s128", | |
| "shape": [ | |
| 128, | |
| 1024 | |
| ] | |
| }, | |
| { | |
| "signature": "s128", | |
| "shape": [ | |
| 128, | |
| 1024 | |
| ] | |
| }, | |
| { | |
| "signature": "s128", | |
| "shape": [ | |
| 128, | |
| 1024 | |
| ] | |
| }, | |
| { | |
| "signature": "s128", | |
| "shape": [ | |
| 128, | |
| 1024 | |
| ] | |
| }, | |
| { | |
| "signature": "s128", | |
| "shape": [ | |
| 128, | |
| 1024 | |
| ] | |
| }, | |
| { | |
| "signature": "s128", | |
| "shape": [ | |
| 128, | |
| 1024 | |
| ] | |
| }, | |
| { | |
| "signature": "s128", | |
| "shape": [ | |
| 128, | |
| 1024 | |
| ] | |
| }, | |
| { | |
| "signature": "s128", | |
| "shape": [ | |
| 128, | |
| 1024 | |
| ] | |
| } | |
| ] | |
| }, | |
| "parity": { | |
| "cpu": { | |
| "calls": 616, | |
| "decisions": 2338, | |
| "agree": 2338, | |
| "agreement": 1.0, | |
| "driftMax": 0.00032, | |
| "driftP99": 0.00017, | |
| "driftP50": 1.7e-05, | |
| "over005": 0, | |
| "msByBucket": { | |
| "128": { | |
| "calls": 8, | |
| "p50": 60.2, | |
| "p95": 61.2 | |
| }, | |
| "160": { | |
| "calls": 321, | |
| "p50": 70.5, | |
| "p95": 72.1 | |
| }, | |
| "512": { | |
| "calls": 287, | |
| "p50": 255.1, | |
| "p95": 263.6 | |
| } | |
| }, | |
| "loadSeconds": 0.61, | |
| "fullyAccelerated": true, | |
| "gate": "pass" | |
| }, | |
| "gpu_fp32": { | |
| "calls": 616, | |
| "decisions": 2338, | |
| "agree": 2338, | |
| "agreement": 1.0, | |
| "driftMax": 0.00032, | |
| "driftP99": 0.00017, | |
| "driftP50": 1.7e-05, | |
| "over005": 0, | |
| "msByBucket": { | |
| "128": { | |
| "calls": 8, | |
| "p50": 22.1, | |
| "p95": 22.6 | |
| }, | |
| "160": { | |
| "calls": 321, | |
| "p50": 25.7, | |
| "p95": 27.2 | |
| }, | |
| "512": { | |
| "calls": 287, | |
| "p50": 79.8, | |
| "p95": 84.1 | |
| } | |
| }, | |
| "loadSeconds": 1.92, | |
| "fullyAccelerated": true, | |
| "gate": "pass" | |
| }, | |
| "gpu_fp16": { | |
| "calls": 616, | |
| "decisions": 2338, | |
| "agree": 2332, | |
| "agreement": 0.9974, | |
| "driftMax": 0.01158, | |
| "driftP99": 0.00692, | |
| "driftP50": 0.001026, | |
| "over005": 0, | |
| "msByBucket": { | |
| "128": { | |
| "calls": 8, | |
| "p50": 16.6, | |
| "p95": 16.9 | |
| }, | |
| "160": { | |
| "calls": 321, | |
| "p50": 22.1, | |
| "p95": 22.6 | |
| }, | |
| "512": { | |
| "calls": 287, | |
| "p50": 50.7, | |
| "p95": 54.7 | |
| } | |
| }, | |
| "loadSeconds": 1.74, | |
| "fullyAccelerated": true, | |
| "gate": "pass" | |
| } | |
| }, | |
| "logitDifferenceFromTorch": { | |
| "cpu": { | |
| "160": 0.000954, | |
| "512": 0.0015 | |
| }, | |
| "gpu_fp32": { | |
| "160": 0.000952, | |
| "512": 0.00151 | |
| }, | |
| "gpu_fp16": { | |
| "160": 0.0228, | |
| "512": 0.0376 | |
| } | |
| }, | |
| "memory": { | |
| "cpuNoCache": { | |
| "loadSeconds": 0.7, | |
| "baseline": { | |
| "rssMb": 194.3, | |
| "footprintMb": 149.4 | |
| }, | |
| "afterLoad": { | |
| "rssMb": 7479.4, | |
| "footprintMb": 6453.6 | |
| }, | |
| "afterEachSignature": { | |
| "s128": { | |
| "rssMb": 7490.3, | |
| "footprintMb": 6464.0 | |
| }, | |
| "s160": { | |
| "rssMb": 7507.9, | |
| "footprintMb": 6481.6 | |
| }, | |
| "s256": { | |
| "rssMb": 7543.6, | |
| "footprintMb": 6517.3 | |
| }, | |
| "s512": { | |
| "rssMb": 7685.7, | |
| "footprintMb": 6659.5 | |
| } | |
| }, | |
| "peakRssMb": 7685.7 | |
| }, | |
| "cpuCacheFirst": { | |
| "loadSeconds": 0.35, | |
| "baseline": { | |
| "rssMb": 194.5, | |
| "footprintMb": 149.6 | |
| }, | |
| "afterLoad": { | |
| "rssMb": 1805.6, | |
| "footprintMb": 776.1 | |
| }, | |
| "afterEachSignature": { | |
| "s128": { | |
| "rssMb": 2968.6, | |
| "footprintMb": 778.0 | |
| }, | |
| "s160": { | |
| "rssMb": 2986.3, | |
| "footprintMb": 795.6 | |
| }, | |
| "s256": { | |
| "rssMb": 3022.0, | |
| "footprintMb": 831.4 | |
| }, | |
| "s512": { | |
| "rssMb": 3164.3, | |
| "footprintMb": 973.8 | |
| } | |
| }, | |
| "peakRssMb": 3164.3 | |
| }, | |
| "cpuCacheWarm": { | |
| "loadSeconds": 0.23, | |
| "baseline": { | |
| "rssMb": 194.5, | |
| "footprintMb": 149.5 | |
| }, | |
| "afterLoad": { | |
| "rssMb": 1783.5, | |
| "footprintMb": 754.0 | |
| }, | |
| "afterEachSignature": { | |
| "s128": { | |
| "rssMb": 2946.5, | |
| "footprintMb": 755.9 | |
| }, | |
| "s160": { | |
| "rssMb": 2964.1, | |
| "footprintMb": 773.5 | |
| }, | |
| "s256": { | |
| "rssMb": 2999.8, | |
| "footprintMb": 809.2 | |
| }, | |
| "s512": { | |
| "rssMb": 3142.1, | |
| "footprintMb": 951.5 | |
| } | |
| }, | |
| "peakRssMb": 3142.1 | |
| } | |
| } | |
| }, | |
| "int8": { | |
| "about": "dynamic-range int8 weights (ai-edge-quantizer dynamic_wi8_afp32: FULLY_CONNECTED and BATCH_MATMUL constants int8 per channel, activations quantized per call); CPU only", | |
| "file": "gliner25_decide_int8.tflite", | |
| "bytes": 422418624, | |
| "buildSeconds": 1.6, | |
| "tableSharing": null, | |
| "tablesFp16": null, | |
| "audit": { | |
| "bytes": 422418624, | |
| "signatures": [ | |
| { | |
| "key": "s128", | |
| "inputs": [ | |
| { | |
| "name": "attention_mask", | |
| "dtype": "float32", | |
| "shape": [ | |
| 1, | |
| 128 | |
| ] | |
| }, | |
| { | |
| "name": "inputs_embeds", | |
| "dtype": "float32", | |
| "shape": [ | |
| 1, | |
| 128, | |
| 1024 | |
| ] | |
| } | |
| ], | |
| "outputs": [ | |
| { | |
| "name": "logits", | |
| "dtype": "float32", | |
| "shape": [ | |
| 1, | |
| 128 | |
| ] | |
| } | |
| ] | |
| }, | |
| { | |
| "key": "s160", | |
| "inputs": [ | |
| { | |
| "name": "attention_mask", | |
| "dtype": "float32", | |
| "shape": [ | |
| 1, | |
| 160 | |
| ] | |
| }, | |
| { | |
| "name": "inputs_embeds", | |
| "dtype": "float32", | |
| "shape": [ | |
| 1, | |
| 160, | |
| 1024 | |
| ] | |
| } | |
| ], | |
| "outputs": [ | |
| { | |
| "name": "logits", | |
| "dtype": "float32", | |
| "shape": [ | |
| 1, | |
| 160 | |
| ] | |
| } | |
| ] | |
| }, | |
| { | |
| "key": "s256", | |
| "inputs": [ | |
| { | |
| "name": "attention_mask", | |
| "dtype": "float32", | |
| "shape": [ | |
| 1, | |
| 256 | |
| ] | |
| }, | |
| { | |
| "name": "inputs_embeds", | |
| "dtype": "float32", | |
| "shape": [ | |
| 1, | |
| 256, | |
| 1024 | |
| ] | |
| } | |
| ], | |
| "outputs": [ | |
| { | |
| "name": "logits", | |
| "dtype": "float32", | |
| "shape": [ | |
| 1, | |
| 256 | |
| ] | |
| } | |
| ] | |
| }, | |
| { | |
| "key": "s512", | |
| "inputs": [ | |
| { | |
| "name": "attention_mask", | |
| "dtype": "float32", | |
| "shape": [ | |
| 1, | |
| 512 | |
| ] | |
| }, | |
| { | |
| "name": "inputs_embeds", | |
| "dtype": "float32", | |
| "shape": [ | |
| 1, | |
| 512, | |
| 1024 | |
| ] | |
| } | |
| ], | |
| "outputs": [ | |
| { | |
| "name": "logits", | |
| "dtype": "float32", | |
| "shape": [ | |
| 1, | |
| 512 | |
| ] | |
| } | |
| ] | |
| } | |
| ], | |
| "operators": { | |
| "s128": 1603, | |
| "s160": 1603, | |
| "s256": 1603, | |
| "s512": 1603 | |
| }, | |
| "opHistogram": { | |
| "s128": { | |
| "ADD": 266, | |
| "BATCH_MATMUL": 96, | |
| "FULLY_CONNECTED": 146, | |
| "GELU": 24, | |
| "MEAN": 98, | |
| "MUL": 220, | |
| "RESHAPE": 341, | |
| "RSQRT": 49, | |
| "SLICE": 96, | |
| "SOFTMAX": 24, | |
| "SQUARED_DIFFERENCE": 49, | |
| "SUB": 50, | |
| "TRANSPOSE": 144 | |
| }, | |
| "s160": { | |
| "ADD": 266, | |
| "BATCH_MATMUL": 96, | |
| "FULLY_CONNECTED": 146, | |
| "GELU": 24, | |
| "MEAN": 98, | |
| "MUL": 220, | |
| "RESHAPE": 341, | |
| "RSQRT": 49, | |
| "SLICE": 96, | |
| "SOFTMAX": 24, | |
| "SQUARED_DIFFERENCE": 49, | |
| "SUB": 50, | |
| "TRANSPOSE": 144 | |
| }, | |
| "s256": { | |
| "ADD": 266, | |
| "BATCH_MATMUL": 96, | |
| "FULLY_CONNECTED": 146, | |
| "GELU": 24, | |
| "MEAN": 98, | |
| "MUL": 220, | |
| "RESHAPE": 341, | |
| "RSQRT": 49, | |
| "SLICE": 96, | |
| "SOFTMAX": 24, | |
| "SQUARED_DIFFERENCE": 49, | |
| "SUB": 50, | |
| "TRANSPOSE": 144 | |
| }, | |
| "s512": { | |
| "ADD": 266, | |
| "BATCH_MATMUL": 96, | |
| "FULLY_CONNECTED": 146, | |
| "GELU": 24, | |
| "MEAN": 98, | |
| "MUL": 220, | |
| "RESHAPE": 341, | |
| "RSQRT": 49, | |
| "SLICE": 96, | |
| "SOFTMAX": 24, | |
| "SQUARED_DIFFERENCE": 49, | |
| "SUB": 50, | |
| "TRANSPOSE": 144 | |
| } | |
| }, | |
| "tensorDtypes": { | |
| "FLOAT32": 7408, | |
| "INT8": 776, | |
| "INT32": 60 | |
| }, | |
| "int64Tensors": 0, | |
| "constantBytesBySignatures": { | |
| "s128+s160+s256+s512": 305375328, | |
| "s512": 50331772, | |
| "s256": 25165948, | |
| "s160": 15728764, | |
| "s128": 12583036 | |
| }, | |
| "activationsWithLeadingDimNot1": [ | |
| { | |
| "signature": "s128", | |
| "shape": [ | |
| 128, | |
| 1024 | |
| ] | |
| }, | |
| { | |
| "signature": "s128", | |
| "shape": [ | |
| 128, | |
| 1024 | |
| ] | |
| }, | |
| { | |
| "signature": "s128", | |
| "shape": [ | |
| 128, | |
| 1024 | |
| ] | |
| }, | |
| { | |
| "signature": "s128", | |
| "shape": [ | |
| 128, | |
| 1024 | |
| ] | |
| }, | |
| { | |
| "signature": "s128", | |
| "shape": [ | |
| 128, | |
| 1024 | |
| ] | |
| }, | |
| { | |
| "signature": "s128", | |
| "shape": [ | |
| 128, | |
| 1024 | |
| ] | |
| }, | |
| { | |
| "signature": "s128", | |
| "shape": [ | |
| 128, | |
| 1024 | |
| ] | |
| }, | |
| { | |
| "signature": "s128", | |
| "shape": [ | |
| 128, | |
| 1024 | |
| ] | |
| } | |
| ] | |
| }, | |
| "parity": { | |
| "cpu": { | |
| "calls": 616, | |
| "decisions": 2338, | |
| "agree": 2306, | |
| "agreement": 0.9863, | |
| "driftMax": 0.05977, | |
| "driftP99": 0.02366, | |
| "driftP50": 0.003493, | |
| "over005": 1, | |
| "msByBucket": { | |
| "128": { | |
| "calls": 8, | |
| "p50": 46.7, | |
| "p95": 47.3 | |
| }, | |
| "160": { | |
| "calls": 321, | |
| "p50": 54.5, | |
| "p95": 55.5 | |
| }, | |
| "512": { | |
| "calls": 287, | |
| "p50": 219.5, | |
| "p95": 224.1 | |
| } | |
| }, | |
| "loadSeconds": 0.86, | |
| "fullyAccelerated": true, | |
| "gate": [ | |
| "probability drift 0.05977" | |
| ] | |
| } | |
| }, | |
| "logitDifferenceFromTorch": { | |
| "cpu": { | |
| "160": 0.083, | |
| "512": 0.107 | |
| } | |
| }, | |
| "memory": { | |
| "cpuNoCache": { | |
| "loadSeconds": 0.9, | |
| "baseline": { | |
| "rssMb": 194.3, | |
| "footprintMb": 149.4 | |
| }, | |
| "afterLoad": { | |
| "rssMb": 1942.5, | |
| "footprintMb": 1489.0 | |
| }, | |
| "afterEachSignature": { | |
| "s128": { | |
| "rssMb": 1954.3, | |
| "footprintMb": 1500.3 | |
| }, | |
| "s160": { | |
| "rssMb": 1971.6, | |
| "footprintMb": 1517.6 | |
| }, | |
| "s256": { | |
| "rssMb": 2008.6, | |
| "footprintMb": 1554.7 | |
| }, | |
| "s512": { | |
| "rssMb": 2152.8, | |
| "footprintMb": 1698.9 | |
| } | |
| }, | |
| "peakRssMb": 2152.8 | |
| }, | |
| "cpuCacheFirst": { | |
| "loadSeconds": 0.33, | |
| "baseline": { | |
| "rssMb": 194.2, | |
| "footprintMb": 149.3 | |
| }, | |
| "afterLoad": { | |
| "rssMb": 768.8, | |
| "footprintMb": 314.3 | |
| }, | |
| "afterEachSignature": { | |
| "s128": { | |
| "rssMb": 1073.3, | |
| "footprintMb": 325.8 | |
| }, | |
| "s160": { | |
| "rssMb": 1090.7, | |
| "footprintMb": 343.1 | |
| }, | |
| "s256": { | |
| "rssMb": 1128.3, | |
| "footprintMb": 380.8 | |
| }, | |
| "s512": { | |
| "rssMb": 1272.8, | |
| "footprintMb": 525.4 | |
| } | |
| }, | |
| "peakRssMb": 1272.8 | |
| }, | |
| "cpuCacheWarm": { | |
| "loadSeconds": 0.13, | |
| "baseline": { | |
| "rssMb": 194.2, | |
| "footprintMb": 149.3 | |
| }, | |
| "afterLoad": { | |
| "rssMb": 473.7, | |
| "footprintMb": 310.0 | |
| }, | |
| "afterEachSignature": { | |
| "s128": { | |
| "rssMb": 778.1, | |
| "footprintMb": 321.3 | |
| }, | |
| "s160": { | |
| "rssMb": 795.6, | |
| "footprintMb": 338.8 | |
| }, | |
| "s256": { | |
| "rssMb": 833.2, | |
| "footprintMb": 376.4 | |
| }, | |
| "s512": { | |
| "rssMb": 977.7, | |
| "footprintMb": 521.0 | |
| } | |
| }, | |
| "peakRssMb": 977.7 | |
| } | |
| } | |
| } | |
| }, | |
| "fp16hw": { | |
| "rule": "k per LayerNorm, the smallest with max |x - mean| * 2^-k <= 32.0 over the FP32 graph's inputs of all 616 recorded calls at their buckets (padding rows included)", | |
| "k": { | |
| "L00.ln1": 0, | |
| "L00.ln2": 3, | |
| "L01.ln1": 0, | |
| "L01.ln2": 4, | |
| "L02.ln1": 0, | |
| "L02.ln2": 4, | |
| "L03.ln1": 0, | |
| "L03.ln2": 3, | |
| "L04.ln1": 0, | |
| "L04.ln2": 3, | |
| "L05.ln1": 0, | |
| "L05.ln2": 5, | |
| "L06.ln1": 0, | |
| "L06.ln2": 6, | |
| "L07.ln1": 0, | |
| "L07.ln2": 6, | |
| "L08.ln1": 0, | |
| "L08.ln2": 6, | |
| "L09.ln1": 0, | |
| "L09.ln2": 6, | |
| "L10.ln1": 0, | |
| "L10.ln2": 5, | |
| "L11.ln1": 0, | |
| "L11.ln2": 4, | |
| "L12.ln1": 0, | |
| "L12.ln2": 5, | |
| "L13.ln1": 0, | |
| "L13.ln2": 4, | |
| "L14.ln1": 0, | |
| "L14.ln2": 4, | |
| "L15.ln1": 0, | |
| "L15.ln2": 3, | |
| "L16.ln1": 0, | |
| "L16.ln2": 3, | |
| "L17.ln1": 0, | |
| "L17.ln2": 3, | |
| "L18.ln1": 0, | |
| "L18.ln2": 4, | |
| "L19.ln1": 0, | |
| "L19.ln2": 4, | |
| "L20.ln1": 0, | |
| "L20.ln2": 2, | |
| "L21.ln1": 0, | |
| "L21.ln2": 2, | |
| "L22.ln1": 0, | |
| "L22.ln2": 1, | |
| "L23.ln1": 1, | |
| "L23.ln2": 1, | |
| "emb": 0 | |
| }, | |
| "maxAbsDeviation": { | |
| "L00.ln1": 13.89, | |
| "L00.ln2": 185.54, | |
| "L01.ln1": 10.56, | |
| "L01.ln2": 305.1, | |
| "L02.ln1": 13.09, | |
| "L02.ln2": 266.54, | |
| "L03.ln1": 10.38, | |
| "L03.ln2": 136.36, | |
| "L04.ln1": 12.7, | |
| "L04.ln2": 147.98, | |
| "L05.ln1": 30.32, | |
| "L05.ln2": 983.72, | |
| "L06.ln1": 24.73, | |
| "L06.ln2": 1231.3, | |
| "L07.ln1": 21.63, | |
| "L07.ln2": 1423.23, | |
| "L08.ln1": 25.53, | |
| "L08.ln2": 1403.17, | |
| "L09.ln1": 23.11, | |
| "L09.ln2": 1092.32, | |
| "L10.ln1": 20.08, | |
| "L10.ln2": 945.09, | |
| "L11.ln1": 12.93, | |
| "L11.ln2": 408.1, | |
| "L12.ln1": 13.14, | |
| "L12.ln2": 603.89, | |
| "L13.ln1": 16.62, | |
| "L13.ln2": 319.06, | |
| "L14.ln1": 20.06, | |
| "L14.ln2": 333.7, | |
| "L15.ln1": 23.33, | |
| "L15.ln2": 203.17, | |
| "L16.ln1": 24.75, | |
| "L16.ln2": 145.38, | |
| "L17.ln1": 22.85, | |
| "L17.ln2": 230.49, | |
| "L18.ln1": 21.41, | |
| "L18.ln2": 419.01, | |
| "L19.ln1": 23.27, | |
| "L19.ln2": 364.46, | |
| "L20.ln1": 23.82, | |
| "L20.ln2": 118.49, | |
| "L21.ln1": 23.09, | |
| "L21.ln2": 88.98, | |
| "L22.ln1": 21.86, | |
| "L22.ln2": 37.04, | |
| "L23.ln1": 37.22, | |
| "L23.ln2": 36.5, | |
| "emb": 1.18 | |
| }, | |
| "prescaled": 25, | |
| "norms": 49 | |
| } | |
| } | |