text-generation-inference
ClickBench
text-generation-inference | ClickBench | |
---|---|---|
29 | 68 | |
7,881 | 571 | |
6.2% | 3.0% | |
9.6 | 9.0 | |
5 days ago | 1 day ago | |
Python | HTML | |
Apache License 2.0 | GNU General Public License v3.0 or later |
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.
text-generation-inference
- FLaNK AI-April 22, 2024
-
Zephyr 141B, a Mixtral 8x22B fine-tune, is now available in Hugging Chat
I wanted to write that TGI inference engine is not Open Source anymore, but they have reverted the license back to Apache 2.0 for the new version TGI v2.0: https://github.com/huggingface/text-generation-inference/rel...
Good news!
- Hugging Face reverts the license back to Apache 2.0
- HuggingFace text-generation-inference is reverting to Apache 2.0 License
- FLaNK Stack 05 Feb 2024
- Is there any open source app to load a model and expose API like OpenAI?
-
AI Code assistant for about 50-70 users
Setting up a server for multiple users is very different from setting up LLM for yourself. A safe bet would be to just use TGI, which supports continuous batching and is very easy to run via Docker on your server. https://github.com/huggingface/text-generation-inference
-
LocalPilot: Open-source GitHub Copilot on your MacBook
Okay, I actually got local co-pilot set up. You will need these 4 things.
1) CodeLlama 13B or another FIM model https://huggingface.co/codellama/CodeLlama-13b-hf. You want "Fill in Middle" models because you're looking at context on both sides of your cursor.
2) HuggingFace llm-ls https://github.com/huggingface/llm-ls A large language mode Language Server (is this making sense yet)
3) HuggingFace inference framework. https://github.com/huggingface/text-generation-inference At least when I tested you couldn't use something like llama.cpp or exllama with the llm-ls, so you need to break out the heavy duty badboy HuggingFace inference server. Just config and run. Now config and run llm-ls.
4) Okay, I mean you need an editor. I just tried nvim, and this was a few weeks ago, so there may be better support. My expereicen was that is was full honest to god copilot. The CodeLlama models are known to be quite good for its size. The FIM part is great. Boilerplace works so much easier with the surrounding context. I'd like to see more models released that can work this way.
-
Mistral 7B Paper on ArXiv
A simple microservice would be https://github.com/huggingface/text-generation-inference .
Works flawlessly in Docker on my Windows machine, which is extremely shocking.
-
best way to serve llama V2 (llama.cpp VS triton VS HF text generation inference)
I am wondering what is the best / most cost-efficient way to serve llama V2. - llama.cpp (is it production ready or just for playing around?) ? - Triton inference server ? - HF text generation inference ?
ClickBench
-
Loading a trillion rows of weather data into TimescaleDB
TimescaleDB primarily serves operational use cases: Developers building products on top of live data, where you are regularly streaming in fresh data, and you often know what many queries look like a priori, because those are powering your live APIs, dashboards, and product experience.
That's different from a data warehouse or many traditional "OLAP" use cases, where you might dump a big dataset statically, and then people will occasionally do ad-hoc queries against it. This is the big weather dataset file sitting on your desktop that you occasionally query while on holidays.
So it's less about "can you store weather data", but what does that use case look like? How are the queries shaped? Are you saving a single dataset for ad-hoc queries across the entire dataset, or continuously streaming in new data, and aging out or de-prioritizing old data?
In most of the products we serve, customers are often interested in recent data in a very granular format ("shallow and wide"), or longer historical queries along a well defined axis ("deep and narrow").
For example, this is where the benefits of TimescaleDB's segmented columnar compression emerges. It optimizes for those queries which are very common in your application, e.g., an IoT application that groups by or selected by deviceID, crypto/fintech analysis based on the ticker symbol, product analytics based on tenantID, etc.
If you look at Clickbench, what most of the queries say are: Scan ALL the data in your database, and GROUP BY one of the 100 columns in the web analytics logs.
- https://github.com/ClickHouse/ClickBench/blob/main/clickhous...
There are almost no time-predicates in the benchmark that Clickhouse created, but perhaps that is not surprising given it was designed for ad-hoc weblog analytics at Yandex.
So yes, Timescale serves many products today that use weather data, but has made different choices than Clickhouse (or things like DuckDB, pg_analytics, etc) to serve those more operational use cases.
-
Variant in Apache Doris 2.1.0: a new data type 8 times faster than JSON for semi-structured data analysis
We tested with 43 Clickbench SQL queries. Queries on the Variant columns are about 10% slower than those on pre-defined static columns, and 8 times faster than those on JSON columns. (For I/O reasons, most cold runs on JSONB data failed with OOM.)
-
Fair Benchmarking Considered Difficult (2018) [pdf]
I have a project dedicated to this topic: https://github.com/ClickHouse/ClickBench
It is important to explain the limitations of a benchmark, provide a methodology, and make it reproducible. It also has to be simple enough, otherwise it will not be realistic to include a large number of participants.
I'm also collecting all database benchmarks I could find: https://github.com/ClickHouse/ClickHouse/issues/22398
- ClickBench – A Benchmark for Analytical DBMS
- FLaNK Stack 05 Feb 2024
-
Why Postgres RDS didn't work for us
Indeed, ClickHouse results were run on an older instance type of the same family and size (c5.4xlarge for ClickHouse and c6a.4xlarge for Timescale), so if anything ClickHouse results are at a slight disadvantage.
This is an open source benchmark - we'd love contributions from Timescale enthusiasts if we missed something: https://github.com/ClickHouse/ClickBench/
-
Show HN: Stanchion – Column-oriented tables in SQLite
Interesting project! Thank you for open sourcing and sharing. Agree that local and embedded analytics are an increasing trend, I see it too.
A couple of questions:
* I’m curious what the difficulties were in the implementation. I suspect it is quite a challenge to implement this support in the current SQLite architecture, and would curious to know which parts were tricky and any design trade-off you were faced with.
* Aside from ease-of-use (install extension, no need for a separate analytical database system), I wonder if there are additional benefits users can anticipate resulting from a single system architecture vs running an embedded OLAP store like DuckDB or clickhouse-local / chdb side-by-side with SQLite? Do you anticipate performance or resource efficiency gains, for instance?
* I am also curious, what the main difficulty with bringing in a separate analytical database is, assuming it natively integrates with SQLite. I may be biased, but I doubt anything can approach the performance of native column-oriented systems, so I'm curious what the tipping point might be for using this extension vs using an embedded OLAP store in practice.
Btw, would love for you or someone in the community to benchmark Stanchion in ClickBench and submit results! (https://github.com/ClickHouse/ClickBench/)
Disclaimer: I work on ClickHouse.
- ClickBench: A Benchmark for Analytical Databases
- DuckDB performance improvements with the latest release
-
DoorDash manages high-availability CockroachDB clusters at scale
interesting. curious if anyone has benchmarked it relative to other dbs. like: https://benchmark.clickhouse.com/
What are some alternatives?
llama-cpp-python - Python bindings for llama.cpp
starrocks - StarRocks, a Linux Foundation project, is a next-generation sub-second MPP OLAP database for full analytics scenarios, including multi-dimensional analytics, real-time analytics, and ad-hoc queries. InfoWorld’s 2023 BOSSIE Award for best open source software.
ollama - Get up and running with Llama 3, Mistral, Gemma, and other large language models.
duckdb - DuckDB is an in-process SQL OLAP Database Management System
exllama - A more memory-efficient rewrite of the HF transformers implementation of Llama for use with quantized weights.
ClickHouse - ClickHouse® is a free analytics DBMS for big data
basaran - Basaran is an open-source alternative to the OpenAI text completion API. It provides a compatible streaming API for your Hugging Face Transformers-based text generation models.
hosts - 🔒 Consolidating and extending hosts files from several well-curated sources. Optionally pick extensions for porn, social media, and other categories.
FlexGen - Running large language models on a single GPU for throughput-oriented scenarios.
TablePlus - TablePlus macOS issue tracker
vllm - A high-throughput and memory-efficient inference and serving engine for LLMs
clickhouse-bulk - Collects many small inserts to ClickHouse and send in big inserts