1
00:00:00,000 --> 00:00:00,000
Hello guys.

2
00:00:00,000 --> 00:00:03,000
So we are going to continue the fine tuning series.

3
00:00:03,000 --> 00:00:08,000
And in this video I'm going to talk about an amazing platform which is called as laminae.

4
00:00:08,000 --> 00:00:15,000
And this laminae platform has almost each and every thing from using basic LM models to probably create

5
00:00:15,000 --> 00:00:22,000
chat bots till fine tuning different, different LM models with your own custom data.

6
00:00:22,000 --> 00:00:26,000
And trust me, uh, if you have seen the Google gamma model, uh, fine tuning right over there, you

7
00:00:26,000 --> 00:00:28,000
found lot of many difficulties.

8
00:00:28,000 --> 00:00:30,000
Lot, many tasks needed to be done.

9
00:00:30,000 --> 00:00:35,000
But this as a platform simplifies the entire fine tuning process.

10
00:00:35,000 --> 00:00:40,000
So let me just go ahead and show you an example how you can perform fine tuning in this.

11
00:00:40,000 --> 00:00:46,000
And obviously in this you will also be getting some amount of free things like with respect to if I

12
00:00:46,000 --> 00:00:48,000
just go ahead and see the pricing.

13
00:00:48,000 --> 00:00:48,000
Right.

14
00:00:48,000 --> 00:00:53,000
So here in the pricing, uh, in the free you can see, see what we can do.

15
00:00:53,000 --> 00:00:55,000
It can see the full model life cycle.

16
00:00:55,000 --> 00:00:56,000
Choose a model run rack.

17
00:00:56,000 --> 00:00:56,000
Try running.

18
00:00:56,000 --> 00:01:03,000
See values used for inferencing compared to a top open source models like llama three, Mistral 2553.

19
00:01:03,000 --> 00:01:10,000
All these models are specifically available 5000 inferences request per month and then limited turning

20
00:01:10,000 --> 00:01:11,000
request.

21
00:01:11,000 --> 00:01:16,000
And you can also run on Llamas Optimized Compute platform using Lora Beft.

22
00:01:17,000 --> 00:01:21,000
Uh, so already we have learned about Lora Beft quantization.

23
00:01:21,000 --> 00:01:21,000
Lora.

24
00:01:21,000 --> 00:01:22,000
Right.

25
00:01:22,000 --> 00:01:26,000
So we'll try to see how we can probably, uh, do the entire fine tuning process.

26
00:01:26,000 --> 00:01:26,000
Okay.

27
00:01:27,000 --> 00:01:29,000
Um, first of all, let's go ahead and log in over here.

28
00:01:29,000 --> 00:01:35,000
So once you probably log in, you'll be able to see that you'll get redirected to app dot a dot AI.

29
00:01:35,000 --> 00:01:40,000
Uh, once you go over here, you can see that out of 5000 total inferences, I have somewhere around

30
00:01:40,000 --> 00:01:42,000
this many number remaining.

31
00:01:42,000 --> 00:01:46,000
So obviously when you can try this much, definitely you'll be able to understand how does the entire

32
00:01:46,000 --> 00:01:48,000
fine tuning process continue?

33
00:01:48,000 --> 00:01:48,000
Right.

34
00:01:48,000 --> 00:01:52,000
And currently this is the API token that they actually provide.

35
00:01:52,000 --> 00:01:53,000
Okay.

36
00:01:53,000 --> 00:01:55,000
Along with this you can go ahead and train your model.

37
00:01:55,000 --> 00:02:00,000
So let's say that this is the model that I have actually trained you know, and here you can see this

38
00:02:00,000 --> 00:02:05,000
model with respect to the custom data set how it is and how was the base model before.

39
00:02:05,000 --> 00:02:05,000
Right.

40
00:02:05,000 --> 00:02:10,000
So all those results also you'll be able to compare, you'll be able to go through the playground and

41
00:02:10,000 --> 00:02:12,000
ask any kind of questions.

42
00:02:12,000 --> 00:02:17,000
Uh, you can probably go ahead and check out the loss, uh, logs with respect to the entire fine tuning

43
00:02:17,000 --> 00:02:19,000
process that also you can probably do, right?

44
00:02:19,000 --> 00:02:24,000
Not only that, here is the entire documentation, which you can also go ahead and follow.

45
00:02:24,000 --> 00:02:30,000
What are the documentation how to use this entire, uh, laminae, uh, using this particular library,

46
00:02:30,000 --> 00:02:31,000
how you can basically start okay.

47
00:02:31,000 --> 00:02:31,000
Okay.

48
00:02:31,000 --> 00:02:38,000
So, uh, this is really, uh, you know, amazing because, uh, completely for free, you should you

49
00:02:38,000 --> 00:02:41,000
you will be definitely able to access all these things.

50
00:02:41,000 --> 00:02:46,000
Okay, so let's start and let me just show you a basic example with respect to fine tuning.

51
00:02:46,000 --> 00:02:51,000
And in this particular example I will go ahead and um, you know, take llama three model.

52
00:02:51,000 --> 00:02:55,000
Uh, and then we'll try to see that how things works over here okay.

53
00:02:55,000 --> 00:03:01,000
So first of all, I will go back to my the same graph line chain.

54
00:03:01,000 --> 00:03:03,000
I'll be using the same requirement.

55
00:03:03,000 --> 00:03:07,000
Uh dot txt over here because I may require all these libraries.

56
00:03:07,000 --> 00:03:08,000
I may not require neo for J.

57
00:03:09,000 --> 00:03:12,000
Uh, but I will just try to show you with respect to that over here.

58
00:03:12,000 --> 00:03:14,000
To start with, we require one model.

59
00:03:14,000 --> 00:03:15,000
Okay.

60
00:03:15,000 --> 00:03:18,000
Which model is that we definitely require.

61
00:03:18,000 --> 00:03:19,000
For that it is nothing but lemon.

62
00:03:19,000 --> 00:03:21,000
So let's go ahead.

63
00:03:21,000 --> 00:03:25,000
And first of all start with the laminate API.

64
00:03:25,000 --> 00:03:26,000
Sorry laminate model itself.

65
00:03:26,000 --> 00:03:30,000
So we can go ahead and probably start doing it okay.

66
00:03:30,000 --> 00:03:34,000
So first of all I will just go ahead and write over here okay.

67
00:03:34,000 --> 00:03:37,000
And I'll show you the step by step how you can basically do.

68
00:03:37,000 --> 00:03:37,000
Okay.

69
00:03:37,000 --> 00:03:44,000
So the first thing that I will be writing is my library which I'm actually going to use okay laminate.

70
00:03:44,000 --> 00:03:46,000
So I will just go over here.

71
00:03:46,000 --> 00:03:51,000
Just go ahead and write pip install minus r requirement dot txt.

72
00:03:51,000 --> 00:03:52,000
Okay.

73
00:03:52,000 --> 00:03:55,000
Okay I'm in different folder okay.

74
00:03:55,000 --> 00:03:56,000
CD dot dot.

75
00:03:56,000 --> 00:03:58,000
And here also I will write CD.

76
00:03:59,000 --> 00:04:01,000
Uh I'm inside my graph lamp chain okay fine.

77
00:04:01,000 --> 00:04:08,000
Uh pip install minus pip install minus r requirement dot txt.

78
00:04:08,000 --> 00:04:11,000
So once you do that, the installation will take place.

79
00:04:11,000 --> 00:04:15,000
I've already done the installation so that it does not take much more time.

80
00:04:15,000 --> 00:04:19,000
Now here I've created one folder called as fine tuning, and here is one fire file that is called as

81
00:04:19,000 --> 00:04:21,000
a fine tuning.py.

82
00:04:21,000 --> 00:04:24,000
I'll be writing my entire code over here itself as we go ahead.

83
00:04:25,000 --> 00:04:30,000
Now, the best thing about using laminae with respect to any kind of models that you go, first of all,

84
00:04:30,000 --> 00:04:31,000
you need to prepare your data.

85
00:04:31,000 --> 00:04:35,000
Okay, one example of preparing a data over here.

86
00:04:35,000 --> 00:04:37,000
I have actually created this.

87
00:04:37,000 --> 00:04:37,000
Okay.

88
00:04:37,000 --> 00:04:41,000
And, uh, I have also checked the documentation.

89
00:04:41,000 --> 00:04:42,000
It was very simple.

90
00:04:42,000 --> 00:04:49,000
So over here you can see I will just try to create a Q and a fine tuning LM model.

91
00:04:49,000 --> 00:04:50,000
I'll use llama three.

92
00:04:50,000 --> 00:04:52,000
But with respect to this particular data okay.

93
00:04:52,000 --> 00:04:58,000
So here you can see in every data you have this in the form of list of key value pairs okay.

94
00:04:58,000 --> 00:05:00,000
So here you have input.

95
00:05:00,000 --> 00:05:02,000
Then you have output okay.

96
00:05:02,000 --> 00:05:05,000
So for this input this is the output for this input.

97
00:05:05,000 --> 00:05:06,000
This is the output.

98
00:05:06,000 --> 00:05:11,000
And if you probably see this it is everything is about laminae itself is the laminae type system similar

99
00:05:11,000 --> 00:05:12,000
to a Python type system.

100
00:05:12,000 --> 00:05:14,000
This is just like f x.

101
00:05:14,000 --> 00:05:17,000
Uh some kind of questions that you have related to laminae.

102
00:05:17,000 --> 00:05:17,000
Right.

103
00:05:17,000 --> 00:05:21,000
So here you have just laminae have a limit on the number of API requests I can make.

104
00:05:21,000 --> 00:05:24,000
Laminae provides each and every user with free tokens up front, right?

105
00:05:24,000 --> 00:05:26,000
So this is the answer.

106
00:05:26,000 --> 00:05:28,000
If I ask this question, this should be the answer.

107
00:05:28,000 --> 00:05:28,000
Like this.

108
00:05:28,000 --> 00:05:33,000
We have created a set of questions and inputs and outputs okay.

109
00:05:33,000 --> 00:05:36,000
So here it actually becomes this particular function.

110
00:05:36,000 --> 00:05:37,000
This function is get data.

111
00:05:37,000 --> 00:05:40,000
And we are creating a variable over here okay.

112
00:05:40,000 --> 00:05:43,000
Now let's go ahead and start this okay.

113
00:05:43,000 --> 00:05:45,000
So this is my data which I have.

114
00:05:45,000 --> 00:05:47,000
And obviously for creating this particular data I don't think so.

115
00:05:47,000 --> 00:05:48,000
You'll require much time.

116
00:05:48,000 --> 00:05:51,000
So first of all I will go ahead and write import laminae okay.

117
00:05:52,000 --> 00:05:57,000
Along with this I'll say hey from laminae import laminae okay.

118
00:05:57,000 --> 00:06:01,000
So there will be this laminae class which I'm going to specifically use.

119
00:06:01,000 --> 00:06:01,000
Okay.

120
00:06:01,000 --> 00:06:06,000
Now I've already told you first to start with we need an API key.

121
00:06:06,000 --> 00:06:13,000
So if you go over here if you go to the eval results or let me just see this okay.

122
00:06:14,000 --> 00:06:15,000
Um.

123
00:06:15,000 --> 00:06:15,000
Mhm.

124
00:06:16,000 --> 00:06:18,000
So here if I go to the account okay.

125
00:06:18,000 --> 00:06:20,000
So this is the API key that I have.

126
00:06:20,000 --> 00:06:22,000
So I'm just going to copy this particular API key.

127
00:06:22,000 --> 00:06:27,000
And I am going to just assign this API key to this particular variable.

128
00:06:27,000 --> 00:06:27,000
Right.

129
00:06:27,000 --> 00:06:35,000
So first of all what I'll write I'll write laminae dot API underscore key is equal to this particular

130
00:06:35,000 --> 00:06:36,000
API key okay.

131
00:06:36,000 --> 00:06:36,000
Okay.

132
00:06:37,000 --> 00:06:39,000
So this is the API key that I'm setting.

133
00:06:39,000 --> 00:06:44,000
You can also store this in your environment variable and call it over here by using AWS dot get env

134
00:06:44,000 --> 00:06:46,000
right that I've already shown you.

135
00:06:46,000 --> 00:06:48,000
Now I will just go ahead and use my LM models.

136
00:06:48,000 --> 00:06:52,000
Now in order to call my LM model, I'll just use this laminae class.

137
00:06:52,000 --> 00:06:56,000
Inside this I will give my model dash name.

138
00:06:56,000 --> 00:07:01,000
Now model underscore name will be in the same way like how you define your model for hugging face.

139
00:07:01,000 --> 00:07:02,000
Right?

140
00:07:02,000 --> 00:07:05,000
So let's say I will go ahead and use this meta llama model.

141
00:07:05,000 --> 00:07:10,000
So it is meta llama slash meta llama okay.

142
00:07:10,000 --> 00:07:13,000
And this we are going to use three.

143
00:07:13,000 --> 00:07:15,000
It is nothing but 8 billion parameter.

144
00:07:15,000 --> 00:07:18,000
And this is nothing but instruct model I'm going to use.

145
00:07:18,000 --> 00:07:21,000
Like this is how the model is available in your hugging face also.

146
00:07:21,000 --> 00:07:21,000
Right?

147
00:07:22,000 --> 00:07:27,000
If you go and see for this particular model, you'll get the entire path Because the best thing is that

148
00:07:27,000 --> 00:07:34,000
this entire thing will be running in the cloud, and it'll be taking care of all the quantization process,

149
00:07:34,000 --> 00:07:35,000
even Laura and Laura.

150
00:07:35,000 --> 00:07:37,000
Every fine tuning will basically happen over there.

151
00:07:37,000 --> 00:07:37,000
Okay.

152
00:07:37,000 --> 00:07:40,000
So this basically becomes my LM model.

153
00:07:40,000 --> 00:07:45,000
Now what I will do I will just go ahead and write data is equal to get underscore data.

154
00:07:45,000 --> 00:07:47,000
Here I get my entire data itself.

155
00:07:47,000 --> 00:07:50,000
Now I'm going to fine tune my entire LM model.

156
00:07:50,000 --> 00:07:54,000
I'll just write LM dot fine tune.

157
00:07:54,000 --> 00:07:56,000
Sorry, LM dot.

158
00:07:56,000 --> 00:07:58,000
There's a function called as tune.

159
00:07:58,000 --> 00:08:06,000
And inside this we will just go ahead and write data or so there is a parameter which is called as data

160
00:08:06,000 --> 00:08:07,000
or data set which is nothing.

161
00:08:07,000 --> 00:08:10,000
But it is assigned to this particular value called as data okay.

162
00:08:11,000 --> 00:08:15,000
Now this is the most simplistic way of fine tuning.

163
00:08:15,000 --> 00:08:19,000
Write this much code will actually help you to get started.

164
00:08:19,000 --> 00:08:22,000
Okay, but we'll not stop here, right?

165
00:08:22,000 --> 00:08:24,000
We will do some more things.

166
00:08:24,000 --> 00:08:26,000
Let's go ahead and add some of the hyper parameters.

167
00:08:26,000 --> 00:08:31,000
Now with respect to the hyper parameters, what all hyper parameters that you have over here okay.

168
00:08:31,000 --> 00:08:37,000
So I'll write down some of the comments so that you will be able to see this common hyper parameters

169
00:08:37,000 --> 00:08:42,000
to tune includes learning rate early stopping max steps and optimizer.

170
00:08:42,000 --> 00:08:45,000
Okay, which are optimizers you can specifically use.

171
00:08:45,000 --> 00:08:50,000
So guys, here are all the common hyper parameters you can probably include for tuning this particular

172
00:08:50,000 --> 00:08:50,000
model.

173
00:08:50,000 --> 00:08:54,000
Let's say learning rate early stopping max steps and optimizer.

174
00:08:54,000 --> 00:08:57,000
So learning rate obviously you can set up some values.

175
00:08:57,000 --> 00:08:57,000
Let me do one thing.

176
00:08:57,000 --> 00:09:00,000
Let me just show you one example how you can go ahead and probably set it up.

177
00:09:00,000 --> 00:09:06,000
So here I'm just going to create my second parameter which is called as fine tune underscore args.

178
00:09:07,000 --> 00:09:09,000
This is what the documentation says.

179
00:09:09,000 --> 00:09:14,000
You need to provide your additional hyper parameter in this particular way in the form of key value

180
00:09:14,000 --> 00:09:14,000
pairs inside this.

181
00:09:14,000 --> 00:09:20,000
So here I'm going to basically go ahead and write learning underscore rate okay.

182
00:09:21,000 --> 00:09:27,000
While we are fine tuning it should be equal to let's say I will go ahead and write 1.4 e to the power

183
00:09:27,000 --> 00:09:28,000
of minus four okay.

184
00:09:28,000 --> 00:09:31,000
Okay, so this will basically be my learning rate.

185
00:09:31,000 --> 00:09:34,000
And uh, I will just use this learning rate and try to probably fine tuning it.

186
00:09:34,000 --> 00:09:40,000
You can still add max steps and optimizer, but I just want to keep it as default values itself.

187
00:09:40,000 --> 00:09:40,000
Okay.

188
00:09:40,000 --> 00:09:45,000
Now let's go ahead and run this so quickly I will go to my fine tuning folder.

189
00:09:45,000 --> 00:09:50,000
I'll say Python fine tune.py.

190
00:09:50,000 --> 00:09:51,000
Okay.

191
00:09:51,000 --> 00:09:56,000
So once I execute this you will be able to see that now this entire thing will be running and it will

192
00:09:56,000 --> 00:09:58,000
first of all load the data set.

193
00:09:58,000 --> 00:09:59,000
See it is uploading the data.

194
00:09:59,000 --> 00:10:02,000
Set data set pairs uploaded the data set.

195
00:10:02,000 --> 00:10:07,000
Is this considering using this feature to train using the same data.

196
00:10:07,000 --> 00:10:09,000
And this is the URL where we will be able to track.

197
00:10:09,000 --> 00:10:15,000
So if I click this URL, it will go ahead and open my landing page okay.

198
00:10:15,000 --> 00:10:17,000
So let's open this.

199
00:10:17,000 --> 00:10:19,000
So oops.

200
00:10:20,000 --> 00:10:23,000
I think it is this page or what.

201
00:10:25,000 --> 00:10:26,000
So 8134.

202
00:10:26,000 --> 00:10:26,000
Right.

203
00:10:26,000 --> 00:10:28,000
So I will open this.

204
00:10:29,000 --> 00:10:31,000
So let's see in this train.

205
00:10:31,000 --> 00:10:34,000
So here you can see 8134 is running.

206
00:10:34,000 --> 00:10:37,000
And before I had tried some more and it got failed.

207
00:10:37,000 --> 00:10:38,000
I stopped it in middle.

208
00:10:38,000 --> 00:10:39,000
Right.

209
00:10:39,000 --> 00:10:41,000
Uh so here you'll also be able to see the logs.

210
00:10:41,000 --> 00:10:43,000
And I've tried multiple things here.

211
00:10:43,000 --> 00:10:48,000
It has been also successful, but I was just exploring multiple things over here.

212
00:10:48,000 --> 00:10:51,000
So here you can see all the logs is basically getting generated.

213
00:10:51,000 --> 00:10:57,000
And if you closely observe the logs here Laura underscore a is there Laura underscore B is there wait

214
00:10:57,000 --> 00:10:57,000
wait.

215
00:10:57,000 --> 00:11:01,000
Related to wait parameters are specifically the all the information is visible over here.

216
00:11:01,000 --> 00:11:02,000
Right.

217
00:11:02,000 --> 00:11:06,000
And this entire training is basically happening and how simple it was.

218
00:11:06,000 --> 00:11:10,000
You just need to put your data in the input and output format.

219
00:11:10,000 --> 00:11:13,000
And just by writing this particular code, you could actually see it, right?

220
00:11:14,000 --> 00:11:15,000
Now let me do one thing.

221
00:11:15,000 --> 00:11:17,000
Let me go back to the completed.

222
00:11:17,000 --> 00:11:21,000
And obviously this process is going to take some amount of time right around five minutes, ten minutes

223
00:11:21,000 --> 00:11:22,000
based on your data set size.

224
00:11:23,000 --> 00:11:28,000
But I have trained with the similar data set over here, and you can see I'm able to get the trained

225
00:11:28,000 --> 00:11:28,000
model.

226
00:11:28,000 --> 00:11:33,000
Okay, but let me talk about some of the important metrics that you'll be seeing after the model is

227
00:11:33,000 --> 00:11:34,000
basically trained.

228
00:11:34,000 --> 00:11:39,000
So the first thing is that it is something related to eval results.

229
00:11:39,000 --> 00:11:44,000
Now what this eval results basically talk about here you can see the model that you have used right.

230
00:11:44,000 --> 00:11:46,000
This model that you see over here.

231
00:11:46,000 --> 00:11:48,000
This is after fine tuning.

232
00:11:48,000 --> 00:11:50,000
And before you had this base model.

233
00:11:50,000 --> 00:11:53,000
This was before fine tuning okay.

234
00:11:54,000 --> 00:11:59,000
Now before fine tuning when When we asked the question, can laminae be used on a regular computer or

235
00:11:59,000 --> 00:12:01,000
do I need specialized hardware or software?

236
00:12:01,000 --> 00:12:02,000
Right.

237
00:12:02,000 --> 00:12:06,000
So here you could probably see if I just go ahead and click this.

238
00:12:06,000 --> 00:12:08,000
This was the answer before.

239
00:12:08,000 --> 00:12:09,000
Right.

240
00:12:09,000 --> 00:12:13,000
But now your answer has much more details when compared to the previous one.

241
00:12:13,000 --> 00:12:13,000
Right.

242
00:12:13,000 --> 00:12:17,000
Similarly, if you go ahead and see the second question, are there any step by step tutorial walkthrough

243
00:12:17,000 --> 00:12:18,000
available in the documentation?

244
00:12:18,000 --> 00:12:22,000
So here if you go ahead and see before the answer used to come like this.

245
00:12:22,000 --> 00:12:25,000
But now you have this entire detailed question.

246
00:12:25,000 --> 00:12:26,000
Detailed answer right.

247
00:12:26,000 --> 00:12:28,000
This is all because of fine tuning, right?

248
00:12:28,000 --> 00:12:33,000
And here similarly you will be able to see what does this cancel underscore job function mean.

249
00:12:33,000 --> 00:12:35,000
Before it was something like this right.

250
00:12:35,000 --> 00:12:36,000
With respect to the base model.

251
00:12:36,000 --> 00:12:39,000
But with the fine tuning model you will be able to see this.

252
00:12:39,000 --> 00:12:40,000
Okay.

253
00:12:40,000 --> 00:12:46,000
So here you will be able to see every kind of new responses with respect to the answers that you are

254
00:12:46,000 --> 00:12:46,000
able to get.

255
00:12:46,000 --> 00:12:47,000
Okay.

256
00:12:47,000 --> 00:12:48,000
The second thing is playground.

257
00:12:48,000 --> 00:12:52,000
If you really want to go ahead and play with this entire models, you can also go ahead and play it

258
00:12:52,000 --> 00:12:53,000
right.

259
00:12:53,000 --> 00:12:57,000
So here only you'll be able to chat and you'll be able to answer ask any kind of questions.

260
00:12:58,000 --> 00:13:02,000
So with respect to this particular playground, you can probably go ahead and ask any questions and

261
00:13:02,000 --> 00:13:04,000
you will be able to test it over here itself.

262
00:13:04,000 --> 00:13:08,000
You'll also be able to get one link, you know, uh, over here with respect to the model.

263
00:13:08,000 --> 00:13:13,000
Now see, this entire thing has got trained right now, if you really want to go ahead and change C

264
00:13:13,000 --> 00:13:17,000
with respect to this argument, the argument that we have used learning rate right here, you are able

265
00:13:17,000 --> 00:13:19,000
to see the eval results right.

266
00:13:19,000 --> 00:13:22,000
This model and your base model response.

267
00:13:22,000 --> 00:13:24,000
And it is like quite a difference over here right.

268
00:13:24,000 --> 00:13:25,000
Which is amazing.

269
00:13:26,000 --> 00:13:31,000
So with respect to this again you can go ahead and ask any question in this playground.

270
00:13:31,000 --> 00:13:33,000
And you can probably ask them okay.

271
00:13:33,000 --> 00:13:35,000
You want to test your model, how the result is.

272
00:13:35,000 --> 00:13:36,000
You can go ahead and test it over here.

273
00:13:36,000 --> 00:13:41,000
Similarly there will be something like you can also create the public link for this model.

274
00:13:41,000 --> 00:13:43,000
So all you have to do is that click on yes confirm.

275
00:13:43,000 --> 00:13:50,000
And uh once this is done, you will be able to share your entire LM model links with anyone so that

276
00:13:50,000 --> 00:13:52,000
they will be able to use it in the chatbot.

277
00:13:52,000 --> 00:13:54,000
Okay, so this will probably take some time.

278
00:13:54,000 --> 00:13:55,000
I've already clicked on.

279
00:13:55,000 --> 00:13:57,000
Yes okay, so let this go.

280
00:13:57,000 --> 00:13:59,000
But overall you'll be able to see that.

281
00:13:59,000 --> 00:14:02,000
Hey how easy, how simple.

282
00:14:02,000 --> 00:14:05,000
It was probably to train this entire model, right?

283
00:14:05,000 --> 00:14:09,000
And uh, here you can see right now I have this public link.

284
00:14:09,000 --> 00:14:10,000
I will copy this over here.

285
00:14:10,000 --> 00:14:11,000
I'll paste it over here.

286
00:14:12,000 --> 00:14:15,000
Now you can go ahead and ask any question and you can go ahead and try it out.

287
00:14:16,000 --> 00:14:20,000
And again it depends on how much quantity of data you are basically being able to train.

288
00:14:20,000 --> 00:14:25,000
But I hope, uh, this gives you a clear idea about the entire fine tuning process and obviously how

289
00:14:25,000 --> 00:14:27,000
simple and easy it was, right?

290
00:14:27,000 --> 00:14:30,000
So yes, go ahead and explore it from your end.

291
00:14:30,000 --> 00:14:37,000
Uh, as, uh, we will be seeing, uh, uh, that this probably I felt is the most easiest way and it

292
00:14:37,000 --> 00:14:42,000
was my duty to probably go ahead with this specific platform and make your work very much easier.

293
00:14:42,000 --> 00:14:45,000
Okay, so yes, this was it for my side.

294
00:14:45,000 --> 00:14:47,000
Uh, I will see you all in the next video.

295
00:14:47,000 --> 00:14:48,000
Thank you.