Post
4198
- 3 replies
Daniel Fox
FlameF0X
AI & ML interests
Pre-training text generator.
(Brother, im 17)
Please don't try to contact me.
Recent Activity
updated
a model
about 6 hours ago
i3-lab/i3-1b-Ethan
published
a model
about 6 hours ago
i3-lab/i3-1b-Ethan
liked
a model
1 day ago
bigscience/bloom-1b1
Organizations
reacted to
daqc's
post with ๐ฅ
18 days ago
Yo, this is neat
replied to
their
post
28 days ago
Hi bro gosh i couldn't find any way to contact youu XD. I want to ask you few things about the i3 models you created. Email me as soon as possible [email protected]
Hi there,
I don't feel comfortable speaking with people privately since I'm only 17 (soon 18).
But if you have questions about the models, I have two repos on GitHub for them:
All the technical details and code are available there.
reacted to
mitkox's
post with ๐ฅ
2 months ago
Post
2792
Say hello to my little friends! I just unboxed this trio of HP Z2 G1a!
Three is always better than one!
3x AMD Ryzen AI Max+ Pro 395
384GB RAM
24TB of RAID storage
Ubuntu 24.04
ROCm 7.0.2
llama cpp, vLLM and Aibrix
Small, cheap GPUs are about to become the Raspberry Pi of edge AI inference. Sprinkle some
Make sure you own your AI. AI in the cloud is not aligned with you; itโs aligned with the company that owns it.
Three is always better than one!
3x AMD Ryzen AI Max+ Pro 395
384GB RAM
24TB of RAID storage
Ubuntu 24.04
ROCm 7.0.2
llama cpp, vLLM and Aibrix
Small, cheap GPUs are about to become the Raspberry Pi of edge AI inference. Sprinkle some
kubectl fairy dust on top, and suddenly it's a high-availability, self-healing, cloud-native, enterprise-grade AI cluster camping in a closet.Make sure you own your AI. AI in the cloud is not aligned with you; itโs aligned with the company that owns it.
- 3 replies
reacted to
andito's
post with โค๏ธ
2 months ago
Post
1904
Finally, our new paper is out! "๐๐ถ๐ป๐ฒ๐ฉ๐ถ๐๐ถ๐ผ๐ป: ๐ข๐ฝ๐ฒ๐ป ๐๐ฎ๐๐ฎ ๐๐ ๐๐น๐น ๐ฌ๐ผ๐ ๐ก๐ฒ๐ฒ๐ฑ"! ๐ฅณ
FineVision: Open Data Is All You Need (2510.17269)
If you've ever trained a VLM, you know this problem: nobody shares their data mixtures. It's a black box, making replicating SOTA work impossible.
We wanted to change that.
FineVision unifies 200 sources into 24 million samples. With 17.3 million images and 9.5 billion answer tokens, it's the largest open resource of its kind.
In the paper, we share how we built it:
๐ finding and cleaning data at scale
๐งน removing excessive duplicates across sources
๐ค decontaminating against 66 public benchmarks
My favorite part is Figure 6 (in the video!). It's our visual diversity analysis. It shows that FineVision isn't just bigger; it's more balanced and conceptually richer than other open datasets.
NVIDIA's Eagle 2 paper highlighted just how critical this visual diversity is, and our results confirm it: models trained on FineVision consistently outperform those trained on any other open dataset on 11 benchmarks!
๐ To celebrate the paper, Iโm also releasing a concatenated and shuffled version of the full dataset! ๐
Itโs ready to stream, so you can start training your own models right away:
from datasets import load_dataset
d = load_dataset("HuggingFaceM4/FineVision_full_shuffled", split="train", streaming=True)
print(next(iter(d)))
A big shoutout to the first authors: Luis Wiedmann and Orr Zohar. They are rockstars!
FineVision: Open Data Is All You Need (2510.17269)
If you've ever trained a VLM, you know this problem: nobody shares their data mixtures. It's a black box, making replicating SOTA work impossible.
We wanted to change that.
FineVision unifies 200 sources into 24 million samples. With 17.3 million images and 9.5 billion answer tokens, it's the largest open resource of its kind.
In the paper, we share how we built it:
๐ finding and cleaning data at scale
๐งน removing excessive duplicates across sources
๐ค decontaminating against 66 public benchmarks
My favorite part is Figure 6 (in the video!). It's our visual diversity analysis. It shows that FineVision isn't just bigger; it's more balanced and conceptually richer than other open datasets.
NVIDIA's Eagle 2 paper highlighted just how critical this visual diversity is, and our results confirm it: models trained on FineVision consistently outperform those trained on any other open dataset on 11 benchmarks!
๐ To celebrate the paper, Iโm also releasing a concatenated and shuffled version of the full dataset! ๐
HuggingFaceM4/FineVision_full_shuffled Itโs ready to stream, so you can start training your own models right away:
from datasets import load_dataset
d = load_dataset("HuggingFaceM4/FineVision_full_shuffled", split="train", streaming=True)
print(next(iter(d)))
A big shoutout to the first authors: Luis Wiedmann and Orr Zohar. They are rockstars!
reacted to
SelmaNajih001's
post with ๐
3 months ago
Post
2292
Finally, I uploaded the model I developed for my masterโs thesis! Given a financial event, it provides explained predictions based on a dataset of past news and central bank speeches.
Try it out here:
SelmaNajih001/StockPredictionExplanation
(Just restart the space and wait a minute)
The dataset used for RAG can be found here:
SelmaNajih001/FinancialNewsAndCentralBanksSpeeches-Summary-Rag
While the dataset used for the training is:
SelmaNajih001/FinancialClassification
I also wrote an article to explain how I've done the training. You can find it here https://huggingface.co/blog/SelmaNajih001/explainable-financial-predictions
Try it out here:
SelmaNajih001/StockPredictionExplanation
(Just restart the space and wait a minute)
The dataset used for RAG can be found here:
SelmaNajih001/FinancialNewsAndCentralBanksSpeeches-Summary-Rag
While the dataset used for the training is:
SelmaNajih001/FinancialClassification
I also wrote an article to explain how I've done the training. You can find it here https://huggingface.co/blog/SelmaNajih001/explainable-financial-predictions
- 2 replies
reacted to
sergiopaniego's
post with ๐ฅ
3 months ago
Post
2938
A few days ago, Thinking Machines Lab released โLoRA Without Regretโ, showing that LoRA can match full fine-tuning performance when configured right.
Naturally, we decided to reproduce the results with TRL and release a guide!
https://huggingface.co/docs/trl/main/en/lora_without_regret
Naturally, we decided to reproduce the results with TRL and release a guide!
https://huggingface.co/docs/trl/main/en/lora_without_regret
reacted to
mlabonne's
post with ๐ฅ
3 months ago
Post
3740
โ๏ธ New drop of tiny task-specific models!
Want to do data extraction, translation, RAG, tool use, or math on a Raspberry Pi? We got you covered! โ
These tiny models were fine-tuned to perform narrow tasks extremely well, making them competitive with much larger models.
You can deploy them today on-device or even on GPUs for big data operations!
LiquidAI/liquid-nanos-68b98d898414dd94d4d5f99a
Want to do data extraction, translation, RAG, tool use, or math on a Raspberry Pi? We got you covered! โ
These tiny models were fine-tuned to perform narrow tasks extremely well, making them competitive with much larger models.
You can deploy them today on-device or even on GPUs for big data operations!
LiquidAI/liquid-nanos-68b98d898414dd94d4d5f99a
- 1 reply
reacted to
mlabonne's
post with ๐ฅ
5 months ago
Post
6880
Liquid just released two 450M and 1.6B param VLMs!
They're super fast and leverage SigLIP2 NaFlex encoders to handle native resolutions without distortion. It's ideal for on-device deployment in constrained environments like phones.
It's available today on Hugging Face, with an inference and a fine-tuning Colab notebooks.
LiquidAI/LFM2-VL-450M
LiquidAI/LFM2-VL-1.6B
They're super fast and leverage SigLIP2 NaFlex encoders to handle native resolutions without distortion. It's ideal for on-device deployment in constrained environments like phones.
It's available today on Hugging Face, with an inference and a fine-tuning Colab notebooks.
LiquidAI/LFM2-VL-450M
LiquidAI/LFM2-VL-1.6B
replied to
their
post
5 months ago
i know, but i spend too much time and money on pre-training the other models
posted
an
update
5 months ago
Post
4063
I am very sad to say that the budget in creating of SnowflakeCore-G1 1b and 7b MoE models ran out and I can't pre-train them anymore.
-
- 7 replies
posted
an
update
5 months ago
Post
508
the training for SnowflakeCore-G1-1B and 7B would be retaken because now I implemented DeepSpeed and management to use two gpus.
replied to
their
post
5 months ago
:D
posted
an
update
5 months ago
Post
271
The development of SnowflakeCore-G1-7B-MoE it getting delay. In the mean time I am working on SnowflakeCore-G1-1B-MoE witch would be a pre-train chatbot.
-
- 1 reply
reacted to
Wauplin's
post with ๐
5 months ago
Post
3292
Say hello to
We are glad to announce a long-awaited quality-of-life improvement: the Hugging Face CLI has been officially renamed from huggingface-cli to hf!
So... why this change?
Typing huggingface-cli constantly gets old fast. More importantly, the CLIโs command structure became messy as new features were added over time (upload, download, cache management, repo management, etc.). Renaming the CLI is a chance to reorganize commands into a clearer, more consistent format.
We decided not to reinvent the wheel and instead follow a well-known CLI pattern: hf <resource> <action>. Isn't
The full rationale, implementation details, and migration notes are in the blog post: https://huggingface.co/blog/hf-cli
hf: a faster, friendlier Hugging Face CLI โจWe are glad to announce a long-awaited quality-of-life improvement: the Hugging Face CLI has been officially renamed from huggingface-cli to hf!
So... why this change?
Typing huggingface-cli constantly gets old fast. More importantly, the CLIโs command structure became messy as new features were added over time (upload, download, cache management, repo management, etc.). Renaming the CLI is a chance to reorganize commands into a clearer, more consistent format.
We decided not to reinvent the wheel and instead follow a well-known CLI pattern: hf <resource> <action>. Isn't
hf auth login easier to type and remember?The full rationale, implementation details, and migration notes are in the blog post: https://huggingface.co/blog/hf-cli
replied to
their
post
5 months ago
update: pre-training the model would need at least 300GB of ram/vram
posted
an
update
5 months ago
Post
2954
The development of SnowflakeCore-G1-7B-MoE. I can't say when it would be publish yet because it's big and it requires a lot of computational power.
-
- 1 reply
posted
an
update
5 months ago
Post
290
I just finished the benchmarks for https://huggingface.co/FlameF0X/SnowflakeCore-G1-Tiny and https://huggingface.co/FlameF0X/SnowflakeCore-G1-Tiny2 in comparation with
openai-community/gpt2 .
reacted to
Jaward's
post with ๐ฅ
5 months ago
Post
3283
Towards batch sizes too small to meter๐ beautiful work! And my personal favorite so far - I adore peak performance at small/nano scale. Everyone deserves to run/train AGI locally:) our data, our god model!
They showed that:
- you can train LLMs (upto 1B params) with as low as batch_size=1. This is unconventional given small batch sizes can lead to unstable/spiky training runs.
- you can have a stable train run with just vanilla SGD(stochastic gradient descent), no momentum required๐คฏ
- small batch sizes are more robust to hyperparameters (i.e no worries with initialization)
- smaller batch sizes outperforms (โbetter per-Flops performanceโ) larger batch sizes.
โWe recommend that practitioners training large models in memory-constrained settings exploit the benefits of small batch sizes rather than trying to emulate the large batch size setting (e.g., through gradient accumulation) typically used in industry.โ
Iโve been doing this for ages - my mantra: all my experiments must scale on my 8gb ram m2 before moving to gpu. IOW I love being gpu poor. Checkout my nanoAI algo repo: https://github.com/Jaykef/ai-algorithms, all notebooks run on memory as low as 8gb ram
They showed that:
- you can train LLMs (upto 1B params) with as low as batch_size=1. This is unconventional given small batch sizes can lead to unstable/spiky training runs.
- you can have a stable train run with just vanilla SGD(stochastic gradient descent), no momentum required๐คฏ
- small batch sizes are more robust to hyperparameters (i.e no worries with initialization)
- smaller batch sizes outperforms (โbetter per-Flops performanceโ) larger batch sizes.
โWe recommend that practitioners training large models in memory-constrained settings exploit the benefits of small batch sizes rather than trying to emulate the large batch size setting (e.g., through gradient accumulation) typically used in industry.โ
Iโve been doing this for ages - my mantra: all my experiments must scale on my 8gb ram m2 before moving to gpu. IOW I love being gpu poor. Checkout my nanoAI algo repo: https://github.com/Jaykef/ai-algorithms, all notebooks run on memory as low as 8gb ram