I would imagine that not running through this codeweavers patch and through rosetta would have better performance.
689 karma · joined December 31, 2015
E-mail: alex %{dot} floresescar %{@gmail}
Curious about Gaussian Splatting.
Most recently started work on splatinit.c It initializes 3D Gaussian Splats from an image and a depth map. It's fast and outputs to a typical 3D Gaussian Splat .ply file.
https://github.com/afloresescarcega/splatinit
I would imagine that not running through this codeweavers patch and through rosetta would have better performance.
https://developer.apple.com/documentation/naturallanguage/nl...
https://twitter.com/officiallogank/status/166393494793189785...
Addressing the model limitations a bit: in the demonstration data that is provided to the base model, we should prevent computed or "looked up" answers.
I've seen some of the demonstration data that people are using to train instruction-tuned models and are being taught to respond by making up answers to solutions it shouldn't try to compute. Btw, the output is wrong.
{ "instruction": "What would be the output of the following JavaScript snippet?", "input": "let area = 6 * 5;\nlet radius = area / 3.14;", "output": "The output of the JavaScript snippet is the radius, which is 1.91." }, [1]
The UI note for now would get us very far but by filtering out demonstrations that retrieve or compute information we can go even farther.
Symbol tuning [2] is addressing the quality of demonstrations but we can take it further by removing retrievals and computations altogether.
Bonus: we can demonstrate how to make it respond so that the user/agent be informed of how to compute or retrieve.
1: https://github.com/sahil280114/codealpaca/commit/0d265112c70...
1:https://jobs.apple.com/en-us/search?location=united-states-U...
We have already documented evidence of the effect of this. In the GPT-4 technical report [1], they reported contamination of humaneval data in the training data.
They did measure against a "non-contaminated" training set but no idea if that can still be trusted.
Why would this matter? We can have seemingly strong benchmarks for containments but measures poorly against new and quarantined information. Classic over fitting.
Another argument is that data being put out there could very much be wrong and the amounts of it amplified by other models. Take a look at this sample of demonstration data for codealpaca [2]. Not only is its output wrong but bad practices like,making up a random computation without it having access to a place to run a calculation, teaches the model these type of responses are ok.
{ "instruction": "What would be the output of the following JavaScript snippet?", "input": "let area = 6 * 5;\nlet radius = area / 3.14;", "output": "The output of the JavaScript snippet is the radius, which is 1.91." }
1: https://cdn.openai.com/papers/gpt-4.pdf 2: https://github.com/sahil280114/codealpaca/commit/0d265112c70...
What we are mostly seeing when it comes to fine-tuning is making a model promptable. Models like LLaMA or the original GPT3 weren't promptable. They were fine-tuned with demonstration data that looks like a prompt input, prompt output.
See below: { "instruction": "What would be the output of the following JavaScript snippet?", "input": "let area = 6 * 5;\nlet radius = area / 3.14;", "output": "The output of the JavaScript snippet is the radius, which is 1.91." }, [1]
Prompt engineering is really just carefully designing what inputs and outputs on a prompt-ready model work best.
I highly recommend skimming this RLHF article and looking for the parts where it talks about demonstration data [2]
1: https://github.com/sahil280114/codealpaca/blob/master/data/c...
I would strongly recommend against fine-tuning over a set of documents as this is a very lossy information system retrieval system. LLMs are not well suited for information retrieval like databases and search engines.
The applications of fine-tuning that we are seeing have a lot of success is making completion models like LLaMA or original GPT3 become prompt-able. In essence, prompt-tuning or instruction-tuning. That is, giving it the ability to respond with a user prompt, llm output chat interface.
Vector databases, for now, are a great way to store mappings of embeddings of documents with the documents themselves for relevant-document information retrieval.
I would highly recommend skimming this RLHF paper for how demonstration data was used to make a model prompt-able [1]. Keep in mind RLHF is another concept all together and we might be seeing a revolution where it might become optional (thanks to LIMA)!
Someone pointed out that some of the answers are using OpenAI's famous "As an AI...". Soooo you can roughly say that RLHF might still have had an impact here through the training data that came from an RLHF model.
But what we are seeing is a revolution in picking quality demonstration data that might make RLHF an optional last step for fine-tuned models.
I'm bullish on a new technique where the quality of instruction-tuned code model output are measured and evaluated automatically by a machine. I'm calling it RL<machine>F
Code Model outputs a test harness, an implementation, and how to run the implementation for a users inputs. Have RLMF evaluate implementations against the synthetic test harness and against hidden user inputs.
1. They don't know what they are saying until they have said it.
2. Your inputs and its outputs help make the next message.
3. LLMs are not suited for information retrieval like databases and search engines.
LLMs excel at reasoning and predicting subsequent text based on given context. Their strength lies in their ability to generate relevant and cohesive responses.
To optimize results, outline clear rules, strategies, or ideas for the LLM to follow. This helps the model craft, revise, or build upon the established context.
Starting with a precise query and introducing rules or constraints incrementally can help steer the model's output in the desired direction.
Avoid zero-shot queries as these can lead to the model generating unexpected or unrelated responses.
Be cautious while seeking pre-calculated or non-derived answers. Some instruction-tuned models might output incorrect solutions, as they are trained to respond to certain queries without proper context or information.
also, this is my biggest gripe no fault of ours of course: don't seek pre-calculated or non-derived answers. I've seen some of the demonstration data that people are using to train instruction-tuned models and are being taught to respond by making up answers to solutions it shouldn't try to compute. Btw, the output is wrong.
{ "instruction": "What would be the output of the following JavaScript snippet?", "input": "let area = 6 * 5;\nlet radius = area / 3.14;", "output": "The output of the JavaScript snippet is the radius, which is 1.91." },
https://github.com/sahil280114/codealpaca/commit/0d265112c70...
Also are they intending to use Samsung's new purchase of eMagine optics for the higher quality microOLED micro displays with no color filter?
If so, are they then just trying to cut reliance on specific types of displays?
Also microLED != microOLED.
I don't think StarCoderBase is instruction-tuned off the bat but would serve as a good starting point for a new technique.
RLHF is fine for things that are hard to measure and evaluate, but code is runnable and testable.
I propose we try Reinforcement Learning Machine Feedback or RLMF.
Prompts and responses are evaluated by how accurate the response evaluated to. We can then train a reward model to help refine StarCoder.
Can you please share more about the merging context across levels? This sounds interesting!
1: "Language Models are Few-Shot Learners" Brown et al. https://arxiv.org/pdf/2005.14165.pdf
It also supports sending it your clipboard contents when you launch it and it parses the words "clipboard". Good for when in a pinch.
I mapped some repetitive prompts to my text replacement on my phone that auto-expand when I type in the input field `*<some shortcut here>`
1: https://www.macstories.net/ios/introducing-s-gpt-a-shortcut-...
Allows you to input data without needing to look at the keyboard.
The positioning of the keyboard wherever your hands and arms rest comfortably, improving ergonomics.
Haptic Feedback
Multi-tasking. Comfortable typing while not disrupting my AR experience
In the GPT-4 technical report, they reported contamination of humaneval data in the training data.
They did measure against a "non-contaminated" training set but no idea if that can still be trusted.