1 4 22

Danny

TheDrunkenSnail

AI & ML interests

None yet

Recent Activity

reacted to eaddario's post with 👍 4 days ago

Squeezing out tensor bits? I have been tinkering with quantization and pruning to reduce model sizes. So far, I've had modest success in producing, on average, 8% smaller versions with negligible loss of quality, and I think further reductions in the 10-15% range are realistic, but I've come across a behaviour I wasn't expecting! Part of the process I'm following consists of quantizing the embedding and output layers aggressively. Since the embedding layer is more about lookup than complex computation, the vectors representing the relative distances between embeddings are usually preserved well enough making this layer fairly robust to quantization. So far, so good. The output layer, on the other hand, maps the final hidden state to the vocabulary logits and therefore, small changes in these logits could lead to a different probability distribution over the vocabulary, resulting in incorrect word predictions, or so I thought. Surprisingly, I'm finding that even at Q2_K the loss of overall capability is minimal. Was this to be expected? or am I missing something? I have published a version with all the test results if you want to give it a try: https://huggingface.co/eaddario/DeepSeek-R1-Distill-Qwen-7B-GGUF I'll upload other models as time allows. Any ideas / clarifications / suggestions are very much welcomed!

liked a model 10 days ago

DavidAU/L3-Stheno-Maid-Blackroot-Grand-HORROR-16.5B-V1.5-STABLE

reacted to kadirnar's post with 👀 24 days ago

Researchers developed Sonic AI enabling precise facial animation from speech cues 🎧 Decouples head/expression control via audio tone analysis + time-aware fusion for natural long-form synthesis

View all activity

Organizations

TheDrunkenSnail's activity

liked a model 10 days ago

DavidAU/L3-Stheno-Maid-Blackroot-Grand-HORROR-16.5B-V1.5-STABLE

Text Generation • Updated Nov 22, 2024 • 25 • 2

liked a model about 1 month ago

unsloth/Mistral-Small-24B-Base-2501

Text Generation • Updated Feb 2 • 1.71k • 3

liked a dataset about 1 month ago

nbeerbower/GreatFirewall-DPO

Viewer • Updated 5 days ago • 492 • 218 • 8

liked a model about 1 month ago

mradermacher/QwQ-LCoT-14B-Conversational-i1-GGUF

Updated Jan 21 • 571 • 1

liked a Space about 2 months ago

12.7k

Open LLM Leaderboard

🏆

Track, rank and evaluate open LLMs and chatbots

liked 2 models 3 months ago

BlouseJury/Mistral-7B-Discord-0.2

Text Generation • Updated Mar 6, 2024 • 81 • 2

mradermacher/Elusive1.2-8B-Model_Stock-i1-GGUF

Updated Dec 18, 2024 • 321 • 2

liked a dataset 3 months ago

nothingiisreal/Reddit-Dirty-And-WritingPrompts

Viewer • Updated Jul 18, 2024 • 393k • 87 • 50

liked a model 3 months ago

Sao10K/I_am_alive_yay

Updated Nov 30, 2024 • 63

liked a dataset 3 months ago

alpindale/two-million-bluesky-posts

Viewer • Updated Nov 28, 2024 • 2.11M • 1.49k • 197

liked a model 4 months ago

DavidAU/Maximizing-Model-Performance-All-Quants-Types-And-Full-Precision-by-Samplers_Parameters

Updated 4 days ago • 90

liked a model 5 months ago

rain1011/pyramid-flow-sd3

Text-to-Video • Updated Oct 30, 2024 • 820

liked a Space 5 months ago

4.33k

OpenGPT 4o

🔥

GPT 4o like bot.

liked a model 5 months ago

TheDrummer/Cydonia-22B-v1

Updated Sep 18, 2024 • 98 • 56

liked 5 models 6 months ago

liked a Space 6 months ago

1.28k

GGUF My Repo

🦙