hack_ml

You will have to send one page at a time, most of this work has to be done via RAG. Adding a large context (like a whole PDF), still does not work that well in my experience.

hack_ml

2 tahun yang laludiscuss

The ablation studies and the dataset can be found here: https://www.zyphra.com/post/building-zyda-2

hack_ml

2 tahun yang laludiscuss

There are some integrations for stuff like https://docs.rapids.ai/visualization :

HoloViews hvPlot Datashader Plotly Bokeh Seaborn Panel PyDeck cuxfilter node RAPIDS

hack_ml

2 tahun yang laludiscuss

There is dask cudf which gets a lot of the way there.

https://docs.rapids.ai/api/dask-cudf/stable/

hack_ml

2 tahun yang laludiscuss

I was conversing with it in Hinglish (A combination of Hindi and English) which folks in Urban India use and it was pretty on point apart from some use of esoteric hindi words but i think with right prompting we can fix that.

hack_ml

2 tahun yang laludiscuss

Nvidia announces Nemotron-4 15B

introduce Nemotron-4 15B, a 15-billion-parameter large multilingual language model trained on 8 trillion text tokens. Nemotron-4 15B demonstrates strong performance when assessed on English, multilingual, and coding tasks: it outperforms all existing similarly-sized open models on 4 out of 7 downstream evaluation areas and achieves competitive performance to the leading open models in the remaining ones. Specifically, Nemotron-4 15B exhibits the best multilingual capabilities of all similarly-sized models, even outperforming models over four times larger and those explicitly specialized for multilingual tasks.

hack_ml

4 tahun yang laludiscuss

Its seamless to accelerate BERTOPIC on GPU's with cuML now with the latest release. (v0.10.0)

Checkout the docs at: https://maartengr.github.io/BERTopic/faq.html#can-i-use-the-...

All you need to do is below

    from bertopic import BERTopic
    from cuml.cluster import HDBSCAN
    from cuml.manifold import UMAP

    # Create instances of GPU-accelerated UMAP and HDBSCAN
    umap_model = UMAP(n_components=5, n_neighbors=15, min_dist=0.0)
    hdbscan_model = HDBSCAN(min_samples=10, gen_min_span_tree=True)

    # Pass the above models to be used in BERTopic
    topic_model = BERTopic(umap_model=umap_model, hdbscan_model=hdbscan_model)
    topics, probs = topic_model.fit_transform(docs)

hack_ml

5 tahun yang laludiscuss

On the SQL front there has been some active work to make that experience better with DASK.

See dask-sql: https://dask-sql.readthedocs.io/en/latest/pages/api.html

hack_ml

5 tahun yang laludiscuss

You can probably use https://github.com/rapidsai/cudf/tree/main/python/dask_cudf a dask wrapper around cuDF.

hack_ml

5 tahun yang laludiscuss

RAPIDS by NVIDIA has an equivalent API open source version of Sckit-Learn https://docs.rapids.ai/api/cuml/stable/ which seems to offer 100x speedup for a lot of these models.

Submissions

GPU Acceleration with Polars

Nemotron-4 15B large multilingual language model trained on 8T tokens

EffVer: Version your code by the effort required to upgrade

7x speed improvement for LLaMA in less than 10 lines of code

Accelerating Topic Modeling on GPUs with Rapids and Bert Models

Dask as a Spark Replacement

19.25x Faster TF-IDF on GPUA with Dask and Rapids than CPUS

Comma Three [video]

comments