SAIDResearch/SAID-LAM-v1

SAIDResearch

LAM

SAID‑LAM‑v1

LAM (Linear Attention Models) — a new family beyond semantic transformers.
SAID‑LAM‑v1 is Linear Attention Memory.

PHILOSOPHY: DETERMINISM OVER PROBABILITY

"The answer IS X. Because I Said so." — At ANY scale

---

Menu

1. Quick-setup
2. MTEB testing (benchmarks)
3. Model Details
4. Performance
5. Drop-in sentence-transformers Replacement
6. Usage
7. API Reference
8. Model Files
9. Citation
10. Links

Model Details

Property	Value
Model Category	LAM (Linear Attention Models) — SAID-LAM-v1: Linear Attention Memory
Parameters	23,848,788
Embedding Dimension	384
Max Context Length	32,768 tokens
Memory Usage	~95 MB
Complexity	O(n) linear — time AND memory
Framework	Pure Rust (Candle) — no PyTorch required
Package Size	~6 MB binary + 92 MB weights (auto-downloaded)
License	Apache 2.0 (weights) / Proprietary (code)

Performance

O(n) Linear Scaling

LAM scales linearly with input length — empirically validated up to 1M words with R²=1.000, with memory growth from ~0 MB at small inputs up to ~15 MB at 1M words:

STS-B Semantic Quality

Spearman r = 0.8181 on the STS-B test set (1,379 sentence pairs):

MTEB LongEmbed Benchmarks

Combined LongEmbed score (SAID-LAM-v1, average over all six tasks): ~91.0%.

Task	Score
LEMBNeedleRetrieval	100.00%
LEMBPasskeyRetrieval	100.00%
LEMBNarrativeQARetrieval	69.93%
LEMBSummScreenFDRetrieval	96.59%
LEMBQMSumRetrieval	85.76%
LEMBWikimQARetrieval	93.98%

LongEmbed SOTA comparison

Task	SAID-LAM-v1 (23M)	Global SOTA
LEMBNeedleRetrieval	100.00%	100.00%
LEMBPasskeyRetrieval	100.00%	100.00%
LEMBNarrativeQARetrieval	69.93%	66.10%
LEMBSummScreenFDRetrieval	96.59%	99.10%
LEMBQMSumRetrieval	85.76%	83.70%
LEMBWikimQARetrieval	93.98%	91.20%

Quick-setup

This quick setup ensures:

• Code is loaded from your pip-installed said-lam package (not a local checkout).
• Weights are loaded from the Hugging Face cache (auto-downloaded on first run).

venv is the standard, recommended way to keep dependencies isolated. You can install globally, but using a virtual environment avoids conflicts.

macOS / Linux (bash, zsh)

cd /path/to/your/project

# 1) Create & activate a clean virtualenv
python3 -m venv .venv
. .venv/bin/activate
python -m pip install --upgrade pip

# 2) Install SAID-LAM (CPU)
pip install said-lam

# 3) Run the quick end-to-end sanity test (CPU)
python said_quick_test.py

Windows (PowerShell)

cd C:\path\to\your\project

# 1) Create & activate a clean virtualenv
py -m venv .venv
.\.venv\Scripts\Activate.ps1
py -m pip install --upgrade pip

# 2) Install SAID-LAM (CPU)
pip install said-lam

# 3) Run the quick end-to-end sanity test (CPU)
py .\said_quick_test.py

If PowerShell blocks activation, run this once (then retry the activate step):

Set-ExecutionPolicy -ExecutionPolicy RemoteSigned -Scope CurrentUser

GPU upgrade (CUDA) uses a separate package. Uninstall CPU first to avoid namespace conflicts (both use import said_lam):

pip uninstall -y said-lam
pip install --upgrade said-lam-gpu

# Run the same quick test on CUDA (if available)
python said_quick_test.py

Windows (PowerShell) equivalent:

pip uninstall -y said-lam
pip install --upgrade said-lam-gpu
py .\said_quick_test.py

MTEB testing (benchmarks)

For full benchmark-style evaluation (STS tasks, LongEmbed retrieval tasks, cache controls, and result JSON export), use mteb_test.py. This is a heavier workflow than said_quick_test.py and is intended for evaluation/benchmarking:

CPU:

. .venv/bin/activate
pip install -r requirements.txt mteb

# CPU smoke (fast coverage)
python mteb_test.py --smoke --device cpu --no-cache --output-dir ./smoke_results_cpu

# Example: run specific tasks
python mteb_test.py --tasks STS12 STS13 --device cpu --no-cache --output-dir ./results_cpu

GPU (CUDA):

. .venv/bin/activate
pip uninstall -y said-lam
pip install --upgrade said-lam-gpu
pip install -r requirements.txt mteb

# GPU smoke (fast coverage)
python mteb_test.py --smoke --device cuda --no-cache --output-dir ./smoke_results_gpu

# Example: run specific tasks
python mteb_test.py --tasks STS12 STS13 --device cuda --no-cache --output-dir ./results_gpu

If no CUDA device is available in the current runtime, the GPU wheel may fall back to CPU automatically.

Drop-in sentence-transformers Replacement

LAM is a drop-in replacement for sentence-transformers. Same API, same output format.

Before (sentence-transformers):

from sentence_transformers import SentenceTransformer

model = SentenceTransformer("all-MiniLM-L6-v2")
embeddings = model.encode(["Hello world", "Semantic search"])
# embeddings.shape == (2, 384), float32, L2-normalized
similarity = embeddings[0] @ embeddings[1]

After (LAM):

from said_lam import LAM

model = LAM("SAIDResearch/SAID-LAM-v1")
embeddings = model.encode(["Hello world", "Semantic search"])
# embeddings.shape == (2, 384), float32, L2-normalized
similarity = embeddings[0] @ embeddings[1]

Same output format, same shapes, same downstream compatibility. Everything that works with sentence-transformers embeddings (FAISS, ChromaDB, Pinecone, numpy dot product) works with LAM embeddings.

Property	sentence-transformers	LAM
Output	(N, 384) ndarray, float32	(N, 384) ndarray, float32
L2-normalized	Yes (default)	Yes (default)
Cosine sim = dot product	Yes	Yes
Max tokens	512	12K (encode) / 32K (SCA)
Complexity	O(n²) attention	O(n) linear
Framework	PyTorch (~2 GB)	Rust (~6 MB)
Memory at 1M tokens (no chunking)	OOM / impractical	~15 MB

Usage

FREE Tier — Embeddings (up to 12K tokens)

from said_lam import LAM

model = LAM("SAIDResearch/SAID-LAM-v1")
embeddings = model.encode(["Hello world", "Semantic search is powerful"])
# embeddings.shape == (2, 384)

# Cosine similarity (L2-normalized by default)
similarity = embeddings[0] @ embeddings[1]
print(f"Similarity: {similarity:.4f}")

BETA SCA (SAID Crystalline Attention) — MTEB testing only

BETA SCA (SAID Crystalline Attention) is activated for MTEB testing only, to enable perfect LongEmbed context retrieval (e.g. LEMBNeedleRetrieval, LEMBPasskeyRetrieval). Use the MTEB evaluation flow; no signup or activation required for benchmarking.

Common Patterns

Similarity Between Texts

Embeddings are L2-normalized — cosine similarity is just a dot product:

emb = model.encode(["The cat sat on the mat", "A kitten rested on the rug"])
similarity = float(emb[0] @ emb[1])
print(f"Similarity: {similarity:.4f}")  # ~0.5761

Batch Similarity Matrix

import numpy as np

queries = ["How is the weather?", "What time is it?"]
candidates = ["Is it raining today?", "Do you have the time?", "Nice shoes"]

emb_q = model.encode(queries)      # (2, 384)
emb_c = model.encode(candidates)   # (3, 384)
sim_matrix = emb_q @ emb_c.T       # (2, 3)

Semantic Search Over a Corpus (FREE Tier)

import numpy as np

corpus = ["Python is a language", "The Eiffel Tower is in Paris",
          "ML uses neural networks", "Speed of light is 299792458 m/s"]
corpus_emb = model.encode(corpus)

query_emb = model.encode(["fastest thing in physics"])
scores = (query_emb @ corpus_emb.T)[0]
ranked = np.argsort(scores)[::-1]
for i in ranked:
    print(f"  {scores[i]:.4f}  {corpus[i]}")

Matryoshka Dimensionality Reduction

emb_128 = model.encode(["Hello world"], output_dim=128)  # (1, 128)
emb_64  = model.encode(["Hello world"], output_dim=64)   # (1, 64)
# Automatically truncated and re-normalized to unit length

Example impact on STS12 (cosine main_score, GPU):

dim	STS12 score	rel. to 384d
384	0.7493	100.0%
256	0.7472	99.7%
128	0.7459	99.6%
64	0.7327	97.8%

Token Limits

• encode(): Up to 12,000 tokens per text. Returns embeddings for your RAG.
• index() + search(): Up to 32,768 tokens per text (MTEB BETA SCA — LongEmbed/MTEB testing group only). SCA streaming — no embeddings, perfect recall.

encode() — returns one embedding per input text, capped at 12K tokens:

# Each text gets one embedding — long texts are chunked at 12K tokens
embeddings = model.encode(["short text", "very long text..."])  # (2, 384)

# Use output_dim for smaller embeddings (Matryoshka)
embeddings = model.encode(["short text", "very long text..."], output_dim=128)  # (2, 128)

Long documents? encode() caps at 12K tokens. For LongEmbed benchmarks (MTEB BETA SCA testing only), index() + search() support up to 32K tokens via SCA.

MTEB Evaluation

One model, one class: use the same LAM with mteb.evaluate() (LAM implements the global MTEB encoder protocol).

from said_lam import LAM
import mteb

model = LAM("SAIDResearch/SAID-LAM-v1")
tasks = mteb.get_tasks(tasks=["LEMBNeedleRetrieval", "LEMBPasskeyRetrieval"])
results = mteb.evaluate(model=model, tasks=tasks)

API Reference

LAM(model_name_or_path, device)

Parameter	Default	Description
model_name_or_path	"SAIDResearch/SAID-LAM-v1"	Hugging Face model ID to load (default: SAIDResearch/SAID-LAM-v1), or a local directory path pointing to the model files
device	None (auto)	Auto-selects CUDA GPU if available, otherwise CPU

Core Methods

Method	Tier	Description
model.encode(sentences, output_dim=None)	FREE+	Encode to embeddings (384, 256, 128, or 64 dims)
model.index(doc_id, text)	MTEB	Index a document for search (benchmarks)
model.search(query, top_k)	MTEB	Retrieve documents by query (benchmarks)
model.truncate_embeddings(emb, dim)	FREE+	Matryoshka truncation (64/128/256)
model.clear()	MTEB	Clear indexed documents (benchmarks)
model.stats()	FREE+	Model statistics

Tier System

Tier	encode()	(SCA)	How to Get	Features
FREE	12K	—	Default	encode() only — embeddings for RAG
MTEB	12K	32K	Auto-detected	SCA for LongEmbed retrieval (benchmarks only)
LICENSED	32K	32K	Coming soon	+ persistent storage + cloud sync
INFINITE	Unlimited	Unlimited	Coming soon	Oracle mode

GPU Support

CPU wheels are installed by default. For GPU acceleration:

# Build from source with CUDA (Linux)
pip install maturin
maturin build --release --features cuda

# Metal (macOS Apple Silicon)
maturin build --release --features metal

Model Files

File	Size	Description
model.safetensors	92 MB	Model weights (SafeTensors format)
config.json	1 KB	Model configuration
tokenizer.json	467 KB	Tokenizer vocabulary
tokenizer_config.json	350 B	Tokenizer settings
vocab.txt	232 KB	WordPiece vocabulary
special_tokens_map.json	112 B	Special token definitions

Citation

@misc{said-lam-v1,
  title={SAID-LAM-v1: Linear Attention Memory},
  author={SAIDResearch},
  year={2026},
  url={https://saidhome.ai},
  note={23.85M parameter embedding model with O(n) linear complexity.
        384-dim embeddings, 32K context window, 100% NIAH recall.
        Distilled from all-MiniLM-L6-v2. Pure Rust (Candle) implementation.}
}

Links

• Organization: SAIDResearch
• PyPI: said-lam
• Distilled From: all-MiniLM-L6-v2
• Framework: Candle (Hugging Face Rust ML)
• Contact: research@saidhome.ai