Upload vidore3_finance_neomme_260m_li: NeoMME-260M late-interaction multi-vector embeddings

README.md

@@ -0,0 +1,160 @@
+---
+pretty_name: vidore3_finance_neomme_260m_li
+license: cc-by-4.0
+tags:
+- multi-vector
+- late-interaction
+- colbert
+- embeddings
+- retrieval
+- image
+---
+
+# vidore3_finance_neomme_260m_li
+
+Multi-vector (late-interaction) embeddings of **ViDoRe finance** (`vidore/finance`), encoded with
+**[Hcompany/NeoMME-260M-Retriever-ST-late](https://huggingface.co/Hcompany/NeoMME-260M-Retriever-ST-late)** at revision `023be2a8ab9d797f5aa76f5bf8b5dde78d819659`.
+
+**Source data:** Hugging Face dataset [`vidore/vidore_v3_finance_en`](https://huggingface.co/datasets/vidore/vidore_v3_finance_en) at revision `7f432c176d82e27546501ad8064a713ac3071809`, configs `corpus` / `queries` / `qrels`, split `test`, loaded with `datasets`. Document, query and qrel ids are the source's own ids,
+unchanged.
+
+Every document is one variable-length set of 128-d vectors; every query is one variable-length
+set of 128-d vectors. Documents and queries are stored at different precisions (fp16 and fp32
+respectively), see [Encoding](#encoding).
+
+## Files
+
+| file | dtype | shape | contents |
+|---|---|---|---|
+| `documents.npy` | float16 (`<f2`) | `[9,608,572, 128]` | every document vector, concatenated document by document (2,345.8 MiB) |
+| `doclens.npy` | int32 | `[2,942]` | vectors per document; `cumsum` gives offsets |
+| `doc_ids.npy` | `<U4` | `[2,942]` | original document ids |
+| `queries.npy` | float32 (`<f4`) | `[1,854, 94, 128]` | query vectors, zero-padded at the end (85.1 MiB) |
+| `query_lens.npy` | int32 | `[1,854]` | true vectors per query, before padding |
+| `queries_ids.npy` | `<U4` | `[1,854]` | original query ids |
+| `qrels.test.tsv` | text | 8,766 rows | TREC qrels, `qid \t 0 \t docid \t relevance`, no header |
+| `gt_top100.tsv` | text | 185,400 rows | exact MaxSim top-100, see below |
+
+All positional indices (`gt_top100.tsv`, and the row order of every `.npy` file) refer to the order
+of `doc_ids.npy` and `queries_ids.npy`. Reordering either file invalidates `gt_top100.tsv`.
+
+## Statistics
+
+| | |
+|---|---|
+| documents | 2,942 |
+| document vectors | 9,608,572 |
+| vectors per document (min / median / mean / max) | 3266 / 3266 / 3266.0 / 3266 |
+| queries | 1,854 |
+| vectors per query (min / median / max) | 18 / 36 / 94 |
+| queries with at least one qrel | 1,854 |
+| qrels rows | 8,766 |
+| embedding dimension | 128 |
+
+## Encoding
+
+| | |
+|---|---|
+| model | [Hcompany/NeoMME-260M-Retriever-ST-late](https://huggingface.co/Hcompany/NeoMME-260M-Retriever-ST-late) |
+| model revision | `023be2a8ab9d797f5aa76f5bf8b5dde78d819659` |
+| library | sentence-transformers 6.0.1 `MultiVectorEncoder` (transformers 5.17.0, torch 2.11.0+cu126) |
+| document compute dtype | float16 (model weights loaded at this dtype for the document pass) |
+| document storage dtype | fp16 |
+| query compute dtype | float32 (model weights loaded at this dtype for the query pass) |
+| query storage dtype | fp32 |
+| normalization | L2, by the model's own `Normalize` module, before the storage cast |
+| document truncation | none (`document_length` unset; model limit 16,384 tokens, longest document here 3,266 vectors, so nothing hit it) |
+| query truncation | none (checkpoint default; `query_length` unset) |
+| document skiplist | none (empty `skiplist_words`): every document token is kept, so doclens is the real token count |
+| document input | page image, one vector per image patch plus layout tokens; processor default resizing |
+| query input | query text, stripped of surrounding whitespace, formatted by the model's own query prompt/template |
+| query vectors | every vector the model emits for the query is kept, including any query-expansion tokens its template adds; `query_lens` counts them all |
+| document padding | none: `documents.npy` holds real vectors only, `sum(doclens) == n_tokens` |
+| query padding | rows at or beyond `query_lens[i]` in `queries.npy[i]` are exactly zero |
+| token_ids | not provided: image-patch vectors have no vocabulary ids (only placeholder ids) |
+
+## Ground truth: `gt_top100.tsv`
+
+Exact brute-force MaxSim top-100 per query over the full corpus, from the vectors in this repo.
+No header; tab-separated `qidx  docidx  rank  score`:
+
+- `qidx`: 0-based row into `queries_ids.npy` / `queries.npy`
+- `docidx`: 0-based position into `doc_ids.npy` / `doclens.npy`
+- `rank`: 1-based, descending score
+- `score`: `sum over the query's query_lens[qidx] vectors of max over the document's vectors of the
+  dot product`, computed in fp32 with the fp16 document vectors upcast to fp32. Expansion vectors
+  are included in the sum. Printed to 6 decimals.
+
+**Self-matches are included.** 55 of 1,854 queries are themselves corpus documents with the same id and retrieve that document (typically at rank 1). This is the raw nearest-neighbour list; BEIR's evaluation drops such pairs (`ignore_identical_ids`), so exclude them before scoring against qrels.
+
+## Retrieval quality
+
+Sanity check of the vectors, not a leaderboard number: exact MaxSim over the full corpus scored
+against `qrels.test.tsv` with ir_measures, with query-id == doc-id pairs dropped as BEIR does.
+
+| nDCG@10 | Recall@100 | MRR@10 | MAP |
+|---|---|---|---|
+| 0.5641 | 0.8664 | 0.6788 | 0.4864 |
+
+## Loading
+
+```python
+import numpy as np
+
+documents = np.load("documents.npy", mmap_mode="r")      # [n_tokens, 128] float16
+doclens = np.load("doclens.npy")                         # [n_docs] int32
+offsets = np.concatenate([[0], np.cumsum(doclens)])
+doc_ids = np.load("doc_ids.npy")                         # [n_docs] str
+
+def document(i):
+    return documents[offsets[i]:offsets[i + 1]]           # [doclens[i], 128]
+
+queries = np.load("queries.npy")                         # [n_queries, 94, 128] float32
+query_lens = np.load("query_lens.npy")                   # [n_queries] int32
+query_ids = np.load("queries_ids.npy")                   # [n_queries] str
+
+def query(j):
+    return queries[j, :query_lens[j]]                     # [query_lens[j], 128]
+
+def maxsim(q, d):
+    return (q @ d.astype(np.float32).T).max(axis=1).sum()
+```
+
+## Validation
+
+Checks run by the exporter on the files exactly as written here:
+
+- ✅ file set — missing=[] extra=[]
+- ✅ documents.npy dtype/shape — <f2 (9608572, 128)
+- ✅ doclens.npy dtype/shape — <i4 (2942,)
+- ✅ doc_ids.npy is a string array — <U4 (2942,)
+- ✅ queries.npy dtype/shape — <f4 (1854, 94, 128)
+- ✅ query_lens.npy dtype/shape — <i4 (1854,)
+- ✅ queries_ids.npy is a string array — <U4 (1854,)
+- ✅ sum(doclens) == n_tokens — 9608572 vs 9608572
+- ✅ no empty documents — min doclen 3266
+- ✅ len(doc_ids) == len(doclens) == corpus size — 2942, 2942, 2942
+- ✅ doc_ids unique and in corpus order
+- ✅ query arrays aligned — 1854, 1854, 1854
+- ✅ queries_ids in dataset order
+- ✅ max(query_lens) == queries.shape[1]
+- ✅ query padding is exactly zero
+- ✅ doc and query dim agree — 128 / 128
+- ✅ document vectors unit-norm (100k sample) — norm range [0.9995, 1.0006]
+- ✅ query vectors unit-norm — norm range [1.000000, 1.000000]
+- ✅ all vectors finite
+- ✅ qrels.test.tsv is 4-col TREC matching the dataset — 8766 rows
+- ✅ gt_top100.tsv has k rows per query — 185400 rows, k=100
+- ✅ gt rows grouped by qidx with ranks 1..k and descending scores
+- ✅ gt indices in range
+- ✅ gt scores reproduce from these files (16 queries, fp32 MaxSim) — max |diff| 1.53e-05
+- ✅ gt top-k is exact (no excluded doc outscores rank k) — max margin -1.05e-04
+- ✅ re-encoded docs match stored rows (4 docs) — lengths match, min token cosine 0.9901
+- ✅ re-encoded queries match stored rows (8 queries) — lengths match, min token cosine 1.000000
+
+## Provenance
+
+| | |
+|---|---|
+| exported | 2026-09-23 |
+| hardware | Tesla V100S-PCIE-32GB |

doc_ids.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fd926ca84c57699e379e53d6324bb6cdbafc9041c1b9f5a310d2a1af07a8ea9f
+size 47200

doclens.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d5bc3f7fdf4198fc09c4ed10cec81f14396f2e538da72e876ae83f7fdbfdfafa
+size 11896

documents.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:54efa6547128e65dfbf5c7d96907fe38841acae1aaa6c15326ff4e1feea47ad8
+size 2459794560

queries.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:69572ce6e0e82dee0e70c57a8ea9006ab96d4216d9ec0a6ff5e2b9fbe40ea525
+size 89229440

queries_ids.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b8770a9d2947a80939496c67a2c7bfd4f996c80b01e459103ed52908941b6717
+size 29792

query_lens.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:84be48276e4f8cbc970db39f8bf7938e6cfd4360fe945ccb5a8898b2ae7b7059
+size 7544