17 3 98

Csaba Kecskemeti PRO

csabakecskemeti

https://devquasar.com/

csabakecskemeti

AI & ML interests

None yet

Recent Activity

updated a model 25 minutes ago

DevQuasar/deepseek-ai.DeepSeek-R1-Zero-bf16

replied to their post about 5 hours ago

I've made an uncensored version of DeepSeek-R1-Distill-Llama-8B with merge. It's passing the "say f***" censor test. Tested with lm-evaluation-harness on standard open llm leaderboard tests + hellaswag. Scores are improved in most. Details on the model card. Model: https://huggingface.co/DevQuasar/DevQuasar-R1-Uncensored-Llama-8B Quants: https://huggingface.co/DevQuasar/DevQuasar-R1-Uncensored-Llama-8B-GGUF

published a Space about 5 hours ago

DevQuasar/Mi50

View all activity

Organizations

csabakecskemeti's activity

updated a model 25 minutes ago

DevQuasar/deepseek-ai.DeepSeek-R1-Zero-bf16

Text Generation • Updated 25 minutes ago

replied to their post about 5 hours ago

Here you can try
https://huggingface.co/spaces/DevQuasar/Mi50

Bust something seems off with my network or with HF everything is very slow.

When llama benched the model I've get 60t/s on the mi50.
Anyway you can try it.

ROCR_VISIBLE_DEVICES=0 build/bin/llama-bench -m ~/Downloads/DevQuasar-R1-Uncensored-Llama-8B.Q8_0.gguf
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 1 ROCm devices:
Device 0: AMD Radeon VII, compute capability 9.0, VMM: no

model	size	params	backend	ngl	test	t/s
llama 8B Q8_0	7.95 GiB	8.03 B	ROCm	99	pp512	416.30 ± 0.07
llama 8B Q8_0	7.95 GiB	8.03 B	ROCm	99	tg128	60.13 ± 0.02

published a Space about 5 hours ago

Running

👁

Mi50

MI50 inference

updated a Space about 5 hours ago

Running

👁

Mi50

MI50 inference

published a model about 18 hours ago

DevQuasar/DevQuasar-R1-Uncensored-Llama-8B

Text Generation • Updated about 18 hours ago • 5 • 1

posted an update about 18 hours ago

Post

917

I've made an uncensored version of DeepSeek-R1-Distill-Llama-8B with merge. It's passing the "say f***" censor test.
Tested with lm-evaluation-harness on standard open llm leaderboard tests + hellaswag. Scores are improved in most. Details on the model card.

Model:
DevQuasar/DevQuasar-R1-Uncensored-Llama-8B
Quants:
DevQuasar/DevQuasar-R1-Uncensored-Llama-8B-GGUF