Hugging Face's logo

TiGa-RCE
/
gte-Qwen2-1.5B-instruct-MLX-Q8

Feature Extraction
MLX
Safetensors
qwen2
embeddings
sentence-similarity
quantization
omlx
q8
8-bit precision
custom_code
Model card Files
xet
 Community
gte-Qwen2-1.5B-instruct-MLX-Q8
File size: 1,918 Bytes
9f3cbea
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
{
  "schema_version": 1,
  "published_repository": "TiGa-RCE/gte-Qwen2-1.5B-instruct-MLX-Q8",
  "family": "gte-Qwen2-1.5B-instruct",
  "variant": "Q8",
  "upstream_repository": "Alibaba-NLP/gte-Qwen2-1.5B-instruct",
  "upstream_revision": "a9af15a6372d7d6b25e9fb07c2ccb9e1fe645644",
  "upstream_revision_evidence": "upstream revision verified at publication time; historical local snapshot metadata was not retained",
  "direct_parent": "TiGa-RCE/gte-Qwen2-1.5B-instruct-MLX-BF16",
  "direct_parent_weight_hashes": [
    {
      "file": "model.safetensors",
      "sha256": "9e9da58bd1371c47a08bc82f58bd29d33a6831094372dd717d68148f061dd11a",
      "bytes": 3552432732
    }
  ],
  "conversion": {
    "method": "Q uniform affine quantization",
    "nominal_bits": 8,
    "group_size": 64,
    "importance_matrix": false,
    "importance_matrix_samples": null,
    "importance_matrix_sequence_length": null,
    "stack": {
      "omlx": "0.5.3",
      "mlx_lm": "0.31.3",
      "mlx": "0.32.0"
    },
    "lossy_parent": false
  },
  "weight_files": [
    {
      "file": "model.safetensors",
      "sha256": "58f2c49af27266301f16ee0724a976aa62c7b1a8ee654d44a509de97c68dff72",
      "bytes": 1887427776
    }
  ],
  "evaluation": {
    "pair_count": 24,
    "top1": 1.0,
    "recall_at_5": 1.0,
    "mrr": 1.0,
    "mean_aligned_embedding_cosine_vs_bf16": 0.9996282458305359,
    "minimum_aligned_embedding_cosine_vs_bf16": 0.9993134140968323,
    "score_rmse_vs_bf16": 0.0022689001634716988,
    "queries_with_rank_change": 0,
    "gate_passed": true,
    "gate_criteria": {
      "top1_delta_min": 0.0,
      "recall_at_5_delta_min": 0.0,
      "mrr_delta_min": -0.01,
      "minimum_aligned_embedding_cosine_min": 0.99,
      "queries_with_rank_change_max": 2
    }
  },
  "collection": "https://huggingface.co/collections/TiGa-RCE/mlx-embedding-quantization-matrix-q-oq-oqe-at-4-6-8-bit-6a68d11afb238d4fe967d70b"
}