trulens
vector
trulens | vector | |
---|---|---|
14 | 97 | |
1,629 | 16,561 | |
7.9% | 1.8% | |
9.8 | 9.9 | |
4 days ago | 3 days ago | |
Jupyter Notebook | Rust | |
MIT License | Mozilla Public License 2.0 |
Stars - the number of stars that a project has on GitHub. Growth - month over month growth in stars.
Activity is a relative number indicating how actively a project is being developed. Recent commits have higher weight than older ones.
For example, an activity of 9.0 indicates that a project is amongst the top 10% of the most actively developed projects that we are tracking.
trulens
-
Why Vector Compression Matters
Retrieval using a single vector is called dense passage retrieval (DPR), because an entire passage (dozens to hundreds of tokens) is encoded as a single vector. ColBERT instead encodes a vector-per-token, where each vector is influenced by surrounding context. This leads to meaningfully better results; for example, here’s ColBERT running on Astra DB compared to DPR using openai-v3-small vectors, compared with TruLens for the Braintrust Coda Help Desk data set. ColBERT easily beats DPR at correctness, context relevance, and groundedness.
- FLaNK AI Weekly 18 March 2024
-
First 15 Open Source Advent projects
12. TruLens by TruEra | Github | tutorial
-
trulens VS agenta - a user suggested alternative
2 projects | 22 Nov 2023
-
How are generative AI companies monitoring their systems in production?
3) Hallucination is probably the biggest problem we solve for. To do evals for hallucination, we typically see our users use a combination of groundedness (does the context support the LLM response) and context relevance (is the retrieved context relevant to the query). There's also a bunch more for the evaluations you mentioned (moderation models, sentiment, usefulness, etc.) and it's pretty easy to add custom evals.
Also - my hot take is that gpt-3.5 is good enough for evals (sometimes better) than gpt-4 if you give the LLM enough instructions on how to do the eval.
website: https://www.trulens.org/
- FLaNK Stack Weekly 28 August 2023
-
[P] TruLens-Eval is an open source project for eval & tracking LLM experiments.
The team at TruEra recently released an open source project for evaluation & tracking of LLM applications called TruLens-Eval. We’ve specifically targeted retrieval-augmented QA as a core use case and so far we’ve seen it used for comparing different models and parameters, prompts, vector-db configurations and query planning strategies. I’d love to get your feedback on it.
- [D] Hardest thing about building with LLMs?
- Stop Evaluating LLMs on Vibes
- OSS library for attribution and interpretation methods for deep nets
vector
-
What is a low/reasonable cost solution for service log storage and querying?
I am thinking about using https://vector.dev/ but would also love opinions on the best deal for lower or reasonable cost storage/querying of logs. Thanks!
-
Docker Log Observability: Analyzing Container Logs in HashiCorp Nomad with Vector, Loki, and Grafana
job "vector" { datacenters = ["dc1"] # system job, runs on all nodes type = "system" group "vector" { count = 1 network { port "api" { to = 8686 } } ephemeral_disk { size = 500 sticky = true } task "vector" { driver = "docker" config { image = "timberio/vector:0.30.0-debian" ports = ["api"] volumes = ["/var/run/docker.sock:/var/run/docker.sock"] } env { VECTOR_CONFIG = "local/vector.toml" VECTOR_REQUIRE_HEALTHY = "false" } resources { cpu = 100 # 100 MHz memory = 100 # 100MB } # template with Vector's configuration template { destination = "local/vector.toml" change_mode = "signal" change_signal = "SIGHUP" # overriding the delimiters to [[ ]] to avoid conflicts with Vector's native templating, which also uses {{ }} left_delimiter = "[[" right_delimiter = "]]" data=<
- FLaNK AI Weekly 18 March 2024
- Vector: A high-performance observability data pipeline
-
Hacks to reduce cloud spend
we are doing something similar with OTEL but we are looking at using https://vector.dev/
-
About reading logs
We don't pull logs, we forward logs to a centralized logging service.
-
Self hosted log paraer
opensearch - amazon fork of Elasticsearch https://opensearch.org/docs/latestif you do this an have distributed log sources you'd use logstash for, bin off logstash and use vector (https://vector.dev/) its better out of the box for SaaS stuff.
-
creating a centralize syslog server with elastic search
I have done something similar in the past: you can send the logs through a centralized syslog servers (I suggest syslog-ng) and from there ingest into ELK. For parsing I am advice to use something like Vector, is a lot more faster than logstash. When you have your logs ingested correctly, you can create your own dashboard in Kibana. If this fit your requirements, no need to install nginx (unless you want to use as reverse proxy for Kibana), php and mysql.
-
Show HN: Homelab Monitoring Setup with Grafana
I think there's nothing currently that combines both logging and metrics into one easy package and visualizes it, but it's also something I would love to have.
Vector[1] would work as the agent, being able to collect both logs and metrics. But the issue would then be storing it. I'm assuming the Elastic Stack might now be able to do both, but it's just to heavy to deal with in a small setup.
A couple of months ago I took a brief look at that when setting up logging for my own homelab (https://pv.wtf/posts/logging-and-the-homelab). Mostly looking at the memory usage to fit it on my synology. Quickwit[2] and Log-Store[3] both come with built in web interfaces that reduce the need for grafana, but neither of them do metrics.
- [1] https://vector.dev
-
Retaining Logs generated by service running in pod.
Log to stdout/stderr and collect your logs with a tool like vector (vector.dev) and send it to something like Grafana Loki.
What are some alternatives?
langfuse - 🪢 Open source LLM engineering platform: Observability, metrics, evals, prompt management, playground, datasets. Integrates with LlamaIndex, Langchain, OpenAI SDK, LiteLLM, and more. 🍊YC W23
graylog - Free and open log management
shapash - 🔅 Shapash: User-friendly Explainability and Interpretability to Develop Reliable and Transparent Machine Learning Models
Fluentd - Fluentd: Unified Logging Layer (project under CNCF)
probability - Probabilistic reasoning and statistical analysis in TensorFlow
agent - Vendor-neutral programmable observability pipelines.
LIME - Tutorial notebooks on explainable Machine Learning with LIME (Original work: https://arxiv.org/abs/1602.04938)
syslog-ng - syslog-ng is an enhanced log daemon, supporting a wide range of input and output methods: syslog, unstructured text, queueing, SQL & NoSQL.
embedchain - Personalizing LLM Responses
OpenSearch - 🔎 Open source distributed and RESTful search engine.
machine_learning_basics - Plain python implementations of basic machine learning algorithms
tracing - Application level tracing for Rust.