GabrielBianconi

6 ms·

31.
We raised $7.3M to build an open-source stack for industrial-grade LLM apps(tensorzero.com)
1 points by GabrielBianconi 1y ago|0 comments
32.
by GabrielBianconi 1y ago
We set up dataset splits and the usual best practices. Of course, if you overdo things, you can still hack benchmarks; our goal isn't to publish SOTA numbers but rather to illustrate results from our methodology. We didn't even tu
33.
by GabrielBianconi 1y ago
Thanks, Sam! I'm excited to see what you guys come up with.
34.
Ask HN: How does the Postgres ecosystem compare to Vitess at 1PB+?
4 points by GabrielBianconi 1y ago|2 comments
35.
by GabrielBianconi 1y ago
With supervised fine-tuning (SFT), you'll often see good results with 100-1000+ datapoints (they can be variations of the same prompt template). If you have more limited data, reinforcement fine-tuning (RFT) can work well in the 10-100
36.
by GabrielBianconi 1y ago
AFAIK, distillation typically refers to tuning on the logits of the larger model, so you wouldn't be able to do that with fine-tuning APIs (OpenAI + Google in our blog post). We fine-tune on the outputs themselves. But broadly speaking
37.
by GabrielBianconi 1y ago
Thanks for the feedback! We chose a set of tasks with different levels of complexity to see how this approach would scale. For LLMs, the "challenge" with NER is not the task itself but the arbitrariness of the labels in the datase
38.
Fine-tuned small LLMs can beat large ones with programmatic data curation(tensorzero.com)
53 points by GabrielBianconi 1y ago|11 comments
39.
by GabrielBianconi 1y ago
[I'm his coworker.] We ran Unsloth ourselves on a GPU-by-the-hour server. We have a notebook in the repository showing how to query historical data and use it with Unsloth. It's a WIP PR that we plan to merge soon: https:/&#
40.
by GabrielBianconi 1y ago
Yeah, I hadn't noticed!
41.
Supervised fine tuning on curated data is reinforcement learning(arxiv.org)
71 points by GabrielBianconi 1y ago|19 comments
42.
by GabrielBianconi 1y ago
TensorZero | https://github.com/tensorzero/tensorzero | Founding Member of Technical Staff | NYC (onsite) | Full-time TensorZero is an open-source stack for industrial-grade LLM applications. It unifies an LLM gateway,
43.
<syntax-highlight>: custom element that uses the CSS Custom Highlight API(andreruffert.github.io)
2 points by GabrielBianconi 1y ago|0 comments
44.
by GabrielBianconi 1y ago
We didn't look into that workflow closely, but you can reproduce our work (code in GitHub) and potentially find some insights! We plan to continue investigating how it works (+ optimize the models and prompts using TensorZero).
45.
by GabrielBianconi 1y ago
They use different prompts depending on the action you're taking. We provided just a sample because our ultimate goal here is to start A/B testing models, optimizing prompts + models, etc. We provide the code to reproduce our work
46.
From NER to Agents: Does Automated Prompt Engineering Scale to Complex Tasks?(tensorzero.com)
1 points by GabrielBianconi 2y ago|0 comments
47.
Can reinforcement learning for LLMs scale beyond math and coding tasks? Probably(arxiv.org)
6 points by GabrielBianconi 2y ago|4 comments
48.
Case Study: Automating Code Changelogs at a Large Bank with LLMs(tensorzero.com)
1 points by GabrielBianconi 2y ago|0 comments
49.
by GabrielBianconi 2y ago
SEEKING FREELANCER TensorZero | https://github.com/tensorzero/tensorzero | Staff Front-end / Design Engineer | Remote or Onsite (NYC) | Full-time or Part-time TensorZero creates a feedback loop for optimizing LLM
50.
by GabrielBianconi 2y ago
TensorZero | https://github.com/tensorzero/tensorzero | Founding Member of Technical Staff | NYC (onsite) | Full-time TensorZero creates a feedback loop for optimizing LLM applications — turning production data into sm
51.
by GabrielBianconi 2y ago
https://www.tensorzero.com/docs/gateway/benchmarks :)
52.
Show HN: TensorZero – open-source data and learning flywheel for LLMs(github.com)
49 points by GabrielBianconi 2y ago|2 comments
53.
by GabrielBianconi 6y ago
SEEKING WORK | MACHINE LEARNING & DATA SCIENCE | REMOTE (NYC) Location: New York (NYC), USA (remote only) Remote: Yes (experienced) Willing to relocate: No Technologies: Python, PyTorch, TensorFlow, Keras, sklearn, PySpark
54.
by GabrielBianconi 6y ago
MACHINE LEARNING & DATA SCIENCE | REMOTE (NYC) | PART-TIME CONTRACT Location: New York (NYC), USA (remote only) Remote: Yes (experienced) Willing to relocate: No Technologies: Python, PyTorch, TensorFlow, Keras, sklearn, P
55.
by GabrielBianconi 6y ago
MACHINE LEARNING & DATA SCIENCE | REMOTE (NYC) | PART-TIME CONTRACT Location: New York (NYC), USA (remote only) Remote: Yes (experienced) Willing to relocate: No Technologies: Python, PyTorch, TensorFlow, Keras, sklearn, P
56.
by GabrielBianconi 6y ago
SEEKING WORK | MACHINE LEARNING & DATA SCIENCE | REMOTE (NYC) Location: New York (NYC), USA (remote only) Remote: Yes (experienced) Willing to relocate: No Technologies: Python, PyTorch, TensorFlow, Keras, sklearn, PySpark
57.
by GabrielBianconi 7y ago
MACHINE LEARNING & DATA SCIENCE | REMOTE (NYC) | PART-TIME CONTRACT Location: New York (NYC), USA (remote only) Remote: Yes (experienced) Willing to relocate: No Technologies: Python, PyTorch, TensorFlow, Keras, sklearn, P
58.
by GabrielBianconi 7y ago
SEEKING WORK | New York (NYC), USA | REMOTE ONLY Senior machine learning engineer (data scientist) with extensive experience in computer vision, NLP, and predictive analytics. * Built products for fast-growing startups (e.g. Scale AI, Fando
59.
Fast R-CNN – Overview(gabrielbianconi.com)
1 points by GabrielBianconi 8y ago|0 comments
60.
R-CNN – Overview(gabrielbianconi.com)
1 points by GabrielBianconi 8y ago|0 comments

More ›