The 2 DeepSeek R1 distilled models available through Ollama are actually very low quality Qwen and Llama models frankensteinged with DeepSeek R1. They cannot be used to judge capabilities of the original DeepSeek R1 in any way, shape, or form.
Ollama and LM Studio cause so much confusion because people simply use their pre-packaged models and not exploring and comparing to what else is available on the market (HuggingFace).
Ollama makes it kind of cumbersome to download straight from Huggingface, unless something changed lately. It doesn't help that they felt the need to store files differently on disk either (inspired by docker, seemingly), making it even harder to share stuff between applications.
LM Studio though allows you to browse and download straight from Huggingface (assuming GGUF), so people could spend more time looking for models, but I don't think many have the interest to do what many of us do, download 10s of models and compare them against each other to find the best one for our use case.
Comments
The 2 DeepSeek R1 distilled models available through Ollama are actually very low quality Qwen and Llama models frankensteinged with DeepSeek R1. They cannot be used to judge capabilities of the original DeepSeek R1 in any way, shape, or form.
Ollama and LM Studio cause so much confusion because people simply use their pre-packaged models and not exploring and comparing to what else is available on the market (HuggingFace).
Ollama makes it kind of cumbersome to download straight from Huggingface, unless something changed lately. It doesn't help that they felt the need to store files differently on disk either (inspired by docker, seemingly), making it even harder to share stuff between applications.
LM Studio though allows you to browse and download straight from Huggingface (assuming GGUF), so people could spend more time looking for models, but I don't think many have the interest to do what many of us do, download 10s of models and compare them against each other to find the best one for our use case.