JDSS/mem0
2
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

docs(rerankers): align reranker config docs with v3 SDK source

Browse Source
This commit is contained in:
kartik-mem0
2026-06-25 11:49:24 +05:30
parent ac296f7534
commit 5345ed5e0b
9 changed files with 55 additions and 64 deletions
Download Patch File Download Diff File Expand all files Collapse all files
docs/components/rerankers/optimization.mdx
+19 -28
View File
@@ -13,10 +13,10 @@ The number of candidates sent to the reranker significantly impacts performance:
```python
# Optimal candidate sizes for different rerankers
config_map = {
"cohere": {"initial_candidates": 100, "top_n": 10},
"sentence_transformer": {"initial_candidates": 50, "top_n": 10},
"huggingface": {"initial_candidates": 30, "top_n": 5},
"llm_reranker": {"initial_candidates": 20, "top_n": 5}
"cohere": {"initial_candidates": 100, "top_k": 10},
"sentence_transformer": {"initial_candidates": 50, "top_k": 10},
"huggingface": {"initial_candidates": 30, "top_k": 5},
"llm_reranker": {"initial_candidates": 20, "top_k": 5}
}
```
@@ -31,7 +31,7 @@ config = {
"config": {
"model": "cross-encoder/ms-marco-MiniLM-L-6-v2",
"batch_size": 16, # Process multiple candidates at once
"top_n": 10
"top_k": 10
}
}
}
@@ -48,7 +48,7 @@ config = {
"provider": "cohere",
"config": {
"model": "rerank-english-v3.0",
"top_n": 10,
"top_k": 10,
"max_chunks_per_doc": 10, # Limit chunk processing
"return_documents": False # Reduce response size
}
@@ -73,7 +73,7 @@ config = {
"model": "cross-encoder/ms-marco-MiniLM-L-6-v2",
"device": "cuda", # Use GPU when available
"batch_size": 32,
"top_n": 10,
"top_k": 10,
"max_length": 512 # Limit input length
}
}
@@ -107,10 +107,9 @@ config = {
"provider": "huggingface",
"config": {
"model": "BAAI/bge-reranker-base",
"use_fp16": True, # Half precision for speed
"max_length": 512,
"batch_size": 8,
"top_n": 10
"top_k": 10
}
}
}
@@ -124,17 +123,11 @@ config = {
"reranker": {
"provider": "llm_reranker",
"config": {
"llm": {
"provider": "openai",
"config": {
"model": "gpt-3.5-turbo", # Faster than gpt-4
"temperature": 0, # Deterministic results
"max_tokens": 500 # Limit response length
}
},
"batch_ranking": True, # Rank multiple at once
"top_n": 5, # Fewer results for faster processing
"timeout": 10 # Request timeout
"provider": "openai",
"model": "gpt-4o-mini", # Faster than gpt-4o
"temperature": 0.0, # Deterministic results
"max_tokens": 100, # Limit response length
"top_k": 5 # Fewer results for faster processing
}
}
}
@@ -201,12 +194,12 @@ class CachedReranker:
### Model Caching
```python
# Pre-load models to avoid initialization overhead
# Set SENTENCE_TRANSFORMERS_HOME env var to control cache location
config = {
"reranker": {
"provider": "sentence_transformer",
"config": {
"model": "cross-encoder/ms-marco-MiniLM-L-6-v2",
"cache_folder": "/path/to/model/cache",
"device": "cuda"
}
}
@@ -218,14 +211,14 @@ config = {
### Async Configuration
```python
import asyncio
from mem0 import Memory
from mem0 import AsyncMemory
async def parallel_search(config, queries, user_id):
memory = Memory.from_config(config)
memory = AsyncMemory.from_config(config)
# Process multiple queries concurrently
tasks = [
memory.search_async(query, user_id=user_id)
memory.search(query, filters={"user_id": user_id})
for query in queries
]
@@ -249,8 +242,7 @@ config = {
"config": {
"device": "cuda",
"model": "cross-encoder/ms-marco-electra-base",
"batch_size": 64, # Larger batch for GPU
"fp16": True # Half precision
"batch_size": 64 # Larger batch for GPU
}
}
}
@@ -268,8 +260,7 @@ config = {
"provider": "sentence_transformer",
"config": {
"device": "cpu",
"model": "cross-encoder/ms-marco-MiniLM-L-6-v2",
"num_workers": 4 # Parallel processing
"model": "cross-encoder/ms-marco-MiniLM-L-6-v2"
}
}
}