diff options
Diffstat (limited to '2023/info/matplotllm-after.md')
| -rw-r--r-- | 2023/info/matplotllm-after.md | 227 |
1 files changed, 227 insertions, 0 deletions
diff --git a/2023/info/matplotllm-after.md b/2023/info/matplotllm-after.md new file mode 100644 index 00000000..2471df19 --- /dev/null +++ b/2023/info/matplotllm-after.md @@ -0,0 +1,227 @@ +<!-- Automatically generated by emacsconf-publish-after-page --> + + +<a name="matplotllm-mainVideo-transcript"></a> +# Transcript + + +[[!template new="1" text="""Introduction""" start="00:00:00.000" video="mainVideo-matplotllm" id="subtitle"]] + +[[!template text="""Hi, my name is Abhinav and I'm going to talk about""" start="00:00:00.000" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""this tool that I've been working on called MatplotLLM.""" start="00:00:03.040" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""MatplotLLM is a natural language interface""" start="00:00:06.200" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""over matplotlib, which is a library I use a lot""" start="00:00:09.520" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""for making visualizations.""" start="00:00:12.480" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""It's a pretty common Python library used a lot everywhere""" start="00:00:14.440" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""where there's need of plotting and graphing.""" start="00:00:18.680" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""I usually use it in reports.""" start="00:00:22.480" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""Whenever I'm writing a report in org mode,""" start="00:00:25.360" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""I tend to write a code block which is in Python.""" start="00:00:27.360" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""And then that code block has usage of matplotlib""" start="00:00:31.560" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""to produce some reports.""" start="00:00:34.080" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""That works really well.""" start="00:00:36.000" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""But at times what happens is""" start="00:00:38.320" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""I have to make a very custom graph, let's say.""" start="00:00:40.000" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""And then while I'm writing a report,""" start="00:00:43.960" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""it's kind of a huge leap of abstraction""" start="00:00:46.920" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""when I'm working on text""" start="00:00:50.680" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""versus going into actual low-level matplotlib code""" start="00:00:51.520" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""to do that graphing.""" start="00:00:54.880" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""So that's something I don't want to do.""" start="00:00:56.240" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""Here's an example.""" start="00:00:59.680" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""This is a graph which is... I think it was made""" start="00:01:00.480" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""like five or six years back.""" start="00:01:04.000" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""And then there are some common things""" start="00:01:05.840" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""like scatter plot here,""" start="00:01:08.400" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""the dots that you can see here scattered.""" start="00:01:09.960" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""Then... But there are a few things which, to do them,""" start="00:01:12.240" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""to make them, you will actually have to go--at least me,""" start="00:01:16.280" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""I have to go to the documentation""" start="00:01:19.160" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""and figure out how to do it. Which is fine,""" start="00:01:20.840" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""but I don't want to do this, you know,""" start="00:01:24.120" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""spend so much time here, when I'm working on""" start="00:01:26.520" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""a tight deadline for a report.""" start="00:01:29.200" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""That's the motivation for this tool.""" start="00:01:32.320" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""This tool basically allows me""" start="00:01:33.920" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""to get rid of the complexity of the library""" start="00:01:35.200" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""by working via an LLM.""" start="00:01:38.480" video="mainVideo-matplotllm" id="subtitle"]] + +[[!template new="1" text="""What is an LLM?""" start="00:01:40.720" video="mainVideo-matplotllm" id="subtitle"]] + +[[!template text="""So an LLM is a large language model.""" start="00:01:40.720" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""These are models which are""" start="00:01:43.400" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""trained to produce text, generate text.""" start="00:01:45.080" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""And just by doing that,""" start="00:01:49.400" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""they actually end up learning a lot of common patterns.""" start="00:01:51.520" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""For example, if you ask a question,""" start="00:01:55.080" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""you can actually get a reasonable response.""" start="00:01:56.800" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""If you ask to write a code for something,""" start="00:01:58.920" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""you'll actually get code""" start="00:02:00.760" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""which can also be very reasonable.""" start="00:02:01.880" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""So this tool is basically a wrapper""" start="00:02:04.760" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""that uses an LLM. For the current version,""" start="00:02:06.600" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""we use GPT-4, which is OpenAI's model.""" start="00:02:11.000" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""It's not open in the sense of open source.""" start="00:02:13.920" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""So that's a problem that it has.""" start="00:02:17.920" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""But for this version, we are going to use that.""" start="00:02:21.120" video="mainVideo-matplotllm" id="subtitle"]] + +[[!template new="1" text="""Using this library""" start="00:02:23.600" video="mainVideo-matplotllm" id="subtitle"]] + +[[!template text="""Using this library is pretty simple.""" start="00:02:23.600" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""You basically require the library""" start="00:02:25.480" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""and then you set up your OpenAI API key here.""" start="00:02:27.400" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""Then you get a code block""" start="00:02:30.720" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""where you can specify the language as `matplotllm`.""" start="00:02:33.360" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""And then what you can do is,""" start="00:02:35.760" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""you can basically describe what you want""" start="00:02:38.280" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""in natural language.""" start="00:02:40.800" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""I'll take this example of this data set.""" start="00:02:41.800" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""It's called the Health and Wealth of Nations.""" start="00:02:45.280" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""I think that was""" start="00:02:48.600" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""the name of a visualization where it was used.""" start="00:02:49.640" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""This is basically life expectancy,""" start="00:02:51.400" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""GDP of various countries starting from 1800.""" start="00:02:53.400" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""I think it goes up to 2000 somewhere.""" start="00:02:59.280" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""So earlier, I would try to write code which reads this CSV""" start="00:03:02.720" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""and then does a lot of matplotlib stuff""" start="00:03:07.480" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""and then finally produces a graph.""" start="00:03:09.840" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""But with this tool, what I'll do is""" start="00:03:11.680" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""I'll just provide instructions in two forms.""" start="00:03:13.880" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""So the first thing I'll do is""" start="00:03:17.680" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""I'll just describe how the data looks like.""" start="00:03:18.880" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""So I'll say data is in a file called `data.csv`,""" start="00:03:21.360" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""which is this file, by the way, on the right.""" start="00:03:29.040" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""It looks like the following.""" start="00:03:33.160" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""I just pasted a few lines from the top, which is enough.""" start="00:03:39.800" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""Since it's a CSV, there's already a structure to it.""" start="00:03:44.360" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""But let's say if you have a log file""" start="00:03:47.120" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""where there's more complexities to be parsed and all,""" start="00:03:50.080" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""that also works out really well.""" start="00:03:53.760" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""You just have to describe how the data looks like""" start="00:03:55.040" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""and the system will figure out how to work with this.""" start="00:03:58.080" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""Now, let's do the plotting. So what I can do is...""" start="00:04:01.160" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""Let's start from a very basic plot""" start="00:04:06.405" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""between life expectancy and GDP per capita.""" start="00:04:09.560" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""I'll just do this.""" start="00:04:11.621" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text=""""Can you make a scatter plot""" start="00:04:13.801" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""for life expectancy and GDP per capita?"""" start="00:04:17.281" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""Now, you can see there are some typos,""" start="00:04:26.400" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""and probably there will be some grammatical mistakes""" start="00:04:29.640" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""also coming through.""" start="00:04:31.720" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""But that's all OK, because the models are supposed to""" start="00:04:32.920" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""handle those kinds of situations really well.""" start="00:04:37.120" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""So I send the request to the model.""" start="00:04:40.560" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""Since it's a large model--GPT-4 is really large--""" start="00:04:43.240" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""it actually takes a lot of time to get the response back.""" start="00:04:47.120" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""So this specific response took 17 seconds,""" start="00:04:50.520" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""which is huge.""" start="00:04:53.360" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""It's not something you would expect""" start="00:04:54.240" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""in a local file running on a computer.""" start="00:04:57.440" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""But I've got what I wanted. Right.""" start="00:04:59.600" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""So there's a scatter plot here, as you can see below,""" start="00:05:01.880" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""which is plotting what I specified it to do,""" start="00:05:04.120" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""though it looks a little dense.""" start="00:05:08.880" video="mainVideo-matplotllm" id="subtitle"]] + +[[!template new="1" text="""Further instructions""" start="00:05:11.701" video="mainVideo-matplotllm" id="subtitle"]] + +[[!template text="""What I can do is""" start="00:05:11.701" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""I can provide further instructions as feedback.""" start="00:05:12.641" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""I try to feed back on this. So I can say,""" start="00:05:16.001" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text=""""Can you only show points where year is the multiple of 50?"""" start="00:05:18.401" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""So since it's starting from 1800, the data points,""" start="00:05:30.600" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""there are too many years,""" start="00:05:33.520" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""so I'll just try to thin them down a little.""" start="00:05:34.720" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""Now what's happening in the background""" start="00:05:37.240" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""is that everything below this last instruction""" start="00:05:40.200" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""is going out as the context to the model""" start="00:05:42.720" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""along with the code that it wrote till now.""" start="00:05:45.720" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""And then this instruction is added on top of it""" start="00:05:47.400" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""so that it basically modifies the code to make it work""" start="00:05:50.080" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""according to this instruction.""" start="00:05:53.080" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""As you can see now, the data points are much fewer.""" start="00:05:55.080" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""This is what I wanted also.""" start="00:05:58.440" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""Let's also do a few more things.""" start="00:06:01.520" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""I want to see the progression through time.""" start="00:06:02.800" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""So maybe I'll do something like, color more recent years""" start="00:06:05.440" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""with a darker shade of...""" start="00:06:13.080" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""Let's change the color map also.""" start="00:06:15.440" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""Now, this again goes back to the model.""" start="00:06:21.720" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""Again, everything below before this line""" start="00:06:24.160" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""is the context along with the current code,""" start="00:06:26.800" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""and then this instruction is going to the model""" start="00:06:29.120" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""to make the changes. So now this should happen, I guess.""" start="00:06:31.800" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""Once this happens. Yeah. So. OK.""" start="00:06:37.040" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""So we have this new color map,""" start="00:06:41.320" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""and there's also this change of color.""" start="00:06:44.600" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""And also there's this range of color from 1800 to 2000,""" start="00:06:46.600" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""which is a nice addition.""" start="00:06:51.720" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""Kind of smart. I didn't expect...""" start="00:06:53.400" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""I didn't exactly ask for it, but it's nice.""" start="00:06:55.840" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""So there's a couple more things.""" start="00:06:58.960" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""Let's make it more minimal. "Let's make it more minimal.""" start="00:07:00.960" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""Can you remove the bounding box?"""" start="00:07:07.760" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""Also, let's annotate a few points.""" start="00:07:17.320" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""So I want to annotate the point""" start="00:07:21.400" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""which has the highest GDP per capita.""" start="00:07:23.720" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text=""""Also annotate the point with highest GDP per capita""" start="00:07:25.840" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""with the country and year."""" start="00:07:33.600" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""So again, forget about the grammar.""" start="00:07:37.000" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""The language model works out well.""" start="00:07:41.600" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""Usually it takes care of""" start="00:07:43.600" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""all those complexities for you.""" start="00:07:46.160" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""This is what we have got after that.""" start="00:07:47.440" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""As you can see, there's the annotation, which is here.""" start="00:07:53.120" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""I think it's still overlapping,""" start="00:07:55.720" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""so probably it could be done better,""" start="00:07:56.680" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""but the box is removed.""" start="00:07:58.560" video="mainVideo-matplotllm" id="subtitle"]] + +[[!template new="1" text="""Room for improvement""" start="00:08:00.160" video="mainVideo-matplotllm" id="subtitle"]] + +[[!template text="""Now, as you can see, the system is...""" start="00:08:00.160" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""You will be able to see this""" start="00:08:03.360" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""that the system is not really robust.""" start="00:08:04.880" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""So the GitHub repository has some examples""" start="00:08:07.480" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""where it fails miserably,""" start="00:08:10.080" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""and you'll actually have to go into the code""" start="00:08:12.120" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""to figure out what's happening.""" start="00:08:13.680" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""But we do expect that to improve slowly,""" start="00:08:15.000" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""because the models are improving greatly in performance.""" start="00:08:17.880" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""This is a very general model.""" start="00:08:21.040" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""This is not even tuned for this use case.""" start="00:08:22.480" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""The other thing is that""" start="00:08:24.480" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""while I was trying to provide feedback,""" start="00:08:26.640" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""I was still using text here all the time,""" start="00:08:29.640" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""but it can be made more natural.""" start="00:08:32.200" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""So, for example, if I have to annotate""" start="00:08:34.560" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""this particular point,""" start="00:08:36.160" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""I actually can just point my cursor to it.""" start="00:08:37.440" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""Emacs has a way to figure out""" start="00:08:42.240" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""where your mouse pointer is.""" start="00:08:44.520" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""And with that, you can actually go back into the code""" start="00:08:45.800" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""and then see which primitive""" start="00:08:49.621" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""is being drawn here in Matplotlib.""" start="00:08:51.961" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""So that there is a way to do that.""" start="00:08:54.481" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""And then, if you do that, then it's really nice to""" start="00:08:55.720" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""just be able to say""" start="00:08:58.440" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""put your cursor here and then say something like,""" start="00:09:01.320" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text=""""Can you make this?""" start="00:09:04.280" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""Can you annotate this point?"""" start="00:09:05.000" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""Because text is, you know... There are limitations to text.""" start="00:09:06.600" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""And if you're producing an image,""" start="00:09:10.720" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""you should be able to do that, too.""" start="00:09:12.480" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""So I do expect that to happen soonish.""" start="00:09:13.960" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""If not, from the model side, the hack that I mentioned""" start="00:09:16.400" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""could be made to work.""" start="00:09:19.840" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""So that will come in in a later version, probably.""" start="00:09:21.360" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""Anyway, so that's the end of my talk.""" start="00:09:24.440" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""You can find more details in the repository link.""" start="00:09:27.600" video="mainVideo-matplotllm" id="subtitle"]] +[[!template text="""Thank you for listening. Goodbye.""" start="00:09:29.760" video="mainVideo-matplotllm" id="subtitle"]] + + + +Captioner: sachac + +Questions or comments? Please e-mail [abhinav@lepisma.xyz](mailto:abhinav@lepisma.xyz?subject=Comment%20for%20EmacsConf%202023%20matplotllm%3A%20MatplotLLM%2C%20iterative%20natural%20language%20data%20visualization%20in%20org-babel) + + +<!-- End of emacsconf-publish-after-page --> |