terafo

Dynamic libraries are a dumpster fire with how they are implemented right now, and I'd really prefer everything to be statically linked. But ideally, I'd like to see exploration of a hybrid solution, where library code is tagged inside a binary, so if the OS detects that multiple applications are using the same version of a library, it's not duplicated in RAM. Such a design would also allow for libraries to be updated if absolutely necessary, either by runtime or some kind of package manager.

This article specifically talks about PC laptops and discusses changes in them.

Having access to the text and being trained on the text are two different things.

There are LLM's that can process 1 million token context window. Amazon Nova 2 for one, even though it's definitely not the highest quality model. You just put whole book in context and make LLM answer questions about it. And given the fact that domain is pretty limited, you can just store KV cache for most popular books on SSD, eliminating quite a bit of cost.

Yes

MLA uses way more flops in order to conserve memory bandwidth, H20 has plenty of memory bandwidth and almost no flops. MLA makes sense on H100/H800, but on H20 GQA-based models are a way better option.

Because you have to do inference distributed between multiple nodes at this point. For prefill because prefill is actually quadratic, but also for memory reasons. KV Cache for 405B at 10M context length would take more than 5 terabytes (at bf16). That's 36 H200 just for KV Cache, but you would need roughly 48 GPUs to serve bf16 version of the model. Generation speed at that setup would be roughly 30 tokens per second, 100k tokens per hour, and you can server only a single user because batching doesn't make sense at these kinds of context lengths. If you pay 3 dollars per hour per GPU, it's $1440 per million tokens cost. For fp8 version the numbers are a bit better: you need only 24 GPUs, generation speed stays roughly the same, so it's only 700 dollars per million tokens. There are architectural modifications that will bring that down significantly, but, nonetheless, it's still really really expensive, but also quite hard to get to work.

Why mention Microsoft twice?

There was. Now second gen of that goes for $15.

Overwhelming majority of flops is indeed spent on matmuls, but softmax disproportionately uses memory bandwidth, so it generally takes much longer than you'd expect from just looking at flops.

t5 is an architecture, t5x is a framework for training models that was created with that architecture in mind, but can be used to train other architectures, including decoder-only ones(there is one in examples).

To quote their official response "If the WSE weren't rectangular, the complexity of power delivery, I/O, mechanical integrity and cooling become much more difficult, to the point of impracticality.".

Not right now.

Because SRAM stopped getting smaller with recent nodes.