Update models.py

models.py

@@ -76,9 +76,9 @@ class OpenAIModelGPT4(BaseTCOModel):
         self.info = gr.Markdown("The cost per input and output tokens values are from OpenAI's [pricing web page](https://openai.com/pricing)", interactive=False, visible=False)
         self.context_length.change(define_cost_per_token, inputs=self.context_length, outputs=[self.input_tokens_cost_per_token, self.output_tokens_cost_per_token])
         
-        self.labor = gr.Number(0, visible=False, 
+        self.labor = gr.Number(1000, visible=False, 
                                 label="($) Labor cost per month", 
-                                info="This is an estimate of the labor cost of the AI engineer in charge of deploying the model",
+                                info="This is an estimate of the labor cost of the AI engineer in charge of deploying the model§/maitenance",
                                 interactive=True
                                 )
 
@@ -100,7 +100,7 @@ class MistralO(BaseTCOModel):
 
     def render(self):
         def define_cost_per_token(context_length):
-            if context_length == "4K":
+            if context_length == "32K":
                 cost_per_1k_input_tokens = 0.0015
                 cost_per_1k_output_tokens = 0.002
             else:
@@ -108,23 +108,23 @@ class MistralO(BaseTCOModel):
                 cost_per_1k_output_tokens = 0.004
             return cost_per_1k_input_tokens, cost_per_1k_output_tokens
 
-        self.context_length = gr.Dropdown(choices=["4K", "16K"], value="4K", interactive=True,
+        self.context_length = gr.Dropdown(choices=["32K"], value="32K", interactive=True,
                                           label="Context size",
                                           visible=False, info="Number of tokens the model considers when processing text")
-        self.input_tokens_cost_per_token = gr.Number(0.0015, visible=False,
+        self.input_tokens_cost_per_token = gr.Number(0.0025, visible=False,
                                            label="($) Price/1K input prompt tokens",
                                            interactive=False
                                            )
-        self.output_tokens_cost_per_token = gr.Number(0.002, visible=False,
+        self.output_tokens_cost_per_token = gr.Number(0.0075, visible=False,
                                            label="($) Price/1K output prompt tokens",
                                            interactive=False
                                            )
-        self.info = gr.Markdown("The cost per input and output tokens values are from OpenAI's [pricing web page](https://openai.com/pricing)", interactive=False, visible=False)
+        self.info = gr.Markdown("The cost per input and output tokens values are from Mistral API", interactive=False, visible=False)
         self.context_length.change(define_cost_per_token, inputs=self.context_length, outputs=[self.input_tokens_cost_per_token, self.output_tokens_cost_per_token])
         
-        self.labor = gr.Number(0, visible=False, 
+        self.labor = gr.Number(1000, visible=False, 
                                 label="($) Labor cost per month", 
-                                info="This is an estimate of the labor cost of the AI engineer in charge of deploying the model",
+                                info="This is an estimate of the labor cost of the AI engineer in charge of deploying the model/maitenance",
                                 interactive=True
                                 )
 
@@ -239,7 +239,7 @@ class DIYLlama2Model(BaseTCOModel):
 class DIYLlama2Model(BaseTCOModel):
     
     def __init__(self):
-        self.set_name("(Deploy yourself) Llama 2/Mistral (et variante 7B)")
+        self.set_name("(Deploy yourself) Llama 2/Mistral (et variante 7B ou 13B)")
         self.set_latency("6s")
         super().__init__()
     
@@ -254,11 +254,11 @@ class DIYLlama2Model(BaseTCOModel):
         self.info = gr.Markdown("The cost per input and output tokens values below are from [these benchmark results](https://www.cursor.so/blog/llama-inference#user-content-fn-llama-paper) that were obtained using the following initial configurations.", 
                                  interactive=False, 
                                  visible=False)
-        self.vm = gr.Textbox(value="2x A100 80GB NVLINK", 
+        self.vm = gr.Textbox(value="2x A6000", 
                               visible=False,
                               label="Instance of VM with GPU",
                               )
-        self.vm_cost_per_hour = gr.Number(4.42, label="Instance cost ($) per hour",
+        self.vm_cost_per_hour = gr.Number(2.58, label="Instance cost ($) per hour",
                                       interactive=False, visible=False)
         self.info_vm = gr.Markdown("This price above is from [CoreWeave's pricing web page](https://www.coreweave.com/gpu-cloud-pricing)", interactive=False, visible=False)
         self.maxed_out = gr.Slider(minimum=1, maximum=100, value=65, step=1, label="Maxed out", info="Estimated average percentage of total GPU memory that is used. The instantaneous value can go from very high when many users are using the service to very low when no one does.", visible=False)
@@ -271,19 +271,19 @@ class DIYLlama2Model(BaseTCOModel):
                                           $TS_{max}$ = Tokens per second when the GPU is maxed out at 100%, <br>
                                           $MO$ = Maxed Out, <br>
                                           """, interactive=False, visible=False)
-        self.input_tokens_cost_per_token = gr.Number(0.00052, visible=False,
+        self.input_tokens_cost_per_token = gr.Number(0.00032, visible=False,
                                            label="($) Price/1K input prompt tokens",
                                            interactive=False
                                            )
-        self.output_tokens_cost_per_token = gr.Number(0.06656, visible=False,
+        self.output_tokens_cost_per_token = gr.Number(0.02656, visible=False,
                                            label="($) Price/1K output prompt tokens",
                                            interactive=False
                                            )
         self.maxed_out.change(on_maxed_out_change, inputs=[self.maxed_out, self.input_tokens_cost_per_token, self.output_tokens_cost_per_token], outputs=[self.input_tokens_cost_per_token, self.output_tokens_cost_per_token])
         
-        self.labor = gr.Number(5000, visible=False, 
-                                label="($) Labor cost per month",
-                                info="This is an estimate of the labor cost of the AI engineer in charge of deploying the model",
+        self.labor = gr.Number(3000, visible=False, 
+                                label="(€) Labor cost per month",
+                                info="This is an estimate of the labor cost of the AI engineer in charge of deploying the model/Maitenance",
                                 interactive=True
                                 )