Shiftrdw
commited on
Commit
·
c57f831
1
Parent(s):
478d83b
update suggest_menu function tool prompt for better output consistency. getting outputs like 'formal dinner', etc. without additional prompt
Browse files
unit2/smolagents/code_agents.ipynb
CHANGED
@@ -39,12 +39,12 @@
|
|
39 |
},
|
40 |
{
|
41 |
"cell_type": "markdown",
|
42 |
-
"source": [
|
43 |
-
"Let's also login to the Hugging Face Hub to have access to the Inference API."
|
44 |
-
],
|
45 |
"metadata": {
|
46 |
"id": "hcfn1TDmVqqg"
|
47 |
-
}
|
|
|
|
|
|
|
48 |
},
|
49 |
{
|
50 |
"cell_type": "code",
|
@@ -725,6 +725,9 @@
|
|
725 |
},
|
726 |
{
|
727 |
"cell_type": "markdown",
|
|
|
|
|
|
|
728 |
"source": [
|
729 |
"When you run this example, the output will **display a trace of the workflow steps being executed**. It will also print the corresponding Python code with the message:\n",
|
730 |
"\n",
|
@@ -736,10 +739,7 @@
|
|
736 |
"```\n",
|
737 |
"\n",
|
738 |
"After a few steps, you'll see the generated playlist that Alfred can use for the party! 🎵"
|
739 |
-
]
|
740 |
-
"metadata": {
|
741 |
-
"id": "waE-prPUVzrh"
|
742 |
-
}
|
743 |
},
|
744 |
{
|
745 |
"cell_type": "markdown",
|
@@ -1021,7 +1021,7 @@
|
|
1021 |
" \"\"\"\n",
|
1022 |
" Suggests a menu based on the occasion.\n",
|
1023 |
" Args:\n",
|
1024 |
-
" occasion: The type of occasion for the party
|
1025 |
" \"\"\"\n",
|
1026 |
" if occasion == \"casual\":\n",
|
1027 |
" return \"Pizza, snacks, and drinks.\"\n",
|
@@ -1039,14 +1039,14 @@
|
|
1039 |
},
|
1040 |
{
|
1041 |
"cell_type": "markdown",
|
|
|
|
|
|
|
1042 |
"source": [
|
1043 |
"The agent will run for a few steps until finding the answer.\n",
|
1044 |
"\n",
|
1045 |
"The menu is ready! 🥗"
|
1046 |
-
]
|
1047 |
-
"metadata": {
|
1048 |
-
"id": "QZmuYtGPWKfO"
|
1049 |
-
}
|
1050 |
},
|
1051 |
{
|
1052 |
"cell_type": "markdown",
|
@@ -1590,15 +1590,15 @@
|
|
1590 |
},
|
1591 |
{
|
1592 |
"cell_type": "markdown",
|
|
|
|
|
|
|
1593 |
"source": [
|
1594 |
"These examples are just the beginning of what you can do with code agents, and we're already starting to see their utility for preparing the party.\n",
|
1595 |
"You can learn more about how to build code agents in the [smolagents documentation](https://huggingface.co/docs/smolagents).\n",
|
1596 |
"\n",
|
1597 |
"`smolagents` specializes in agents that write and execute Python code snippets, offering sandboxed execution for security. It supports both open-source and proprietary language models, making it adaptable to various development environments."
|
1598 |
-
]
|
1599 |
-
"metadata": {
|
1600 |
-
"id": "T7pLiWCcWSsU"
|
1601 |
-
}
|
1602 |
},
|
1603 |
{
|
1604 |
"cell_type": "markdown",
|
@@ -3555,12 +3555,12 @@
|
|
3555 |
},
|
3556 |
{
|
3557 |
"cell_type": "markdown",
|
3558 |
-
"source": [
|
3559 |
-
"To download the agent again, use the code below:"
|
3560 |
-
],
|
3561 |
"metadata": {
|
3562 |
"id": "3iPLK1IQWguH"
|
3563 |
-
}
|
|
|
|
|
|
|
3564 |
},
|
3565 |
{
|
3566 |
"cell_type": "code",
|
@@ -4539,14 +4539,14 @@
|
|
4539 |
},
|
4540 |
{
|
4541 |
"cell_type": "markdown",
|
|
|
|
|
|
|
4542 |
"source": [
|
4543 |
"What's also exciting is that shared agents are directly available as Hugging Face Spaces, allowing you to interact with them in real-time. You can explore other agents [here](https://huggingface.co/spaces/davidberenstein1957/smolagents-and-tools).\n",
|
4544 |
"\n",
|
4545 |
"For example, the _AlfredAgent_ is available [here](https://huggingface.co/spaces/sergiopaniego/AlfredAgent)."
|
4546 |
-
]
|
4547 |
-
"metadata": {
|
4548 |
-
"id": "qz-vwxu9Wobn"
|
4549 |
-
}
|
4550 |
},
|
4551 |
{
|
4552 |
"cell_type": "markdown",
|
@@ -4580,14 +4580,14 @@
|
|
4580 |
},
|
4581 |
{
|
4582 |
"cell_type": "markdown",
|
|
|
|
|
|
|
4583 |
"source": [
|
4584 |
"Next, Alfred has already created an account on Langfuse and has his API keys ready. If you haven’t done so yet, you can sign up for Langfuse Cloud [here](https://cloud.langfuse.com/) or explore [alternatives](https://huggingface.co/docs/smolagents/tutorials/inspect_runs). \n",
|
4585 |
"\n",
|
4586 |
"Once you have your API keys, they need to be properly configured as follows:"
|
4587 |
-
]
|
4588 |
-
"metadata": {
|
4589 |
-
"id": "IVBwrNgbXPS-"
|
4590 |
-
}
|
4591 |
},
|
4592 |
{
|
4593 |
"cell_type": "code",
|
@@ -4612,12 +4612,12 @@
|
|
4612 |
},
|
4613 |
{
|
4614 |
"cell_type": "markdown",
|
4615 |
-
"source": [
|
4616 |
-
"Finally, Alfred is ready to initialize the `SmolagentsInstrumentor` and start tracking his agent's performance. "
|
4617 |
-
],
|
4618 |
"metadata": {
|
4619 |
"id": "ZIRRj-PJXRGH"
|
4620 |
-
}
|
|
|
|
|
|
|
4621 |
},
|
4622 |
{
|
4623 |
"cell_type": "code",
|
@@ -4641,12 +4641,12 @@
|
|
4641 |
},
|
4642 |
{
|
4643 |
"cell_type": "markdown",
|
4644 |
-
"source": [
|
4645 |
-
"Alfred is now connected 🔌! The runs from `smolagents` are being logged in Langfuse, giving him full visibility into the agent's behavior. With this setup, he's ready to revisit previous runs and refine his Party Preparator Agent even further. "
|
4646 |
-
],
|
4647 |
"metadata": {
|
4648 |
"id": "B0oLBr_qXSjA"
|
4649 |
-
}
|
|
|
|
|
|
|
4650 |
},
|
4651 |
{
|
4652 |
"cell_type": "code",
|
@@ -5212,14 +5212,14 @@
|
|
5212 |
},
|
5213 |
{
|
5214 |
"cell_type": "markdown",
|
|
|
|
|
|
|
5215 |
"source": [
|
5216 |
"Alfred can now access this logs [here](https://cloud.langfuse.com/project/cm7bq0abj025rad078ak3luwi/traces/995fc019255528e4f48cf6770b0ce27b?timestamp=2025-02-19T10%3A28%3A36.929Z) to review and analyze them. \n",
|
5217 |
"\n",
|
5218 |
"Meanwhile, the [suggested playlist](https://open.spotify.com/playlist/0gZMMHjuxMrrybQ7wTMTpw) sets the perfect vibe for the party preparations. Cool, right? 🎶\n"
|
5219 |
-
]
|
5220 |
-
"metadata": {
|
5221 |
-
"id": "KBpqTBK1XVLz"
|
5222 |
-
}
|
5223 |
}
|
5224 |
],
|
5225 |
"metadata": {
|
@@ -10631,4 +10631,4 @@
|
|
10631 |
},
|
10632 |
"nbformat": 4,
|
10633 |
"nbformat_minor": 0
|
10634 |
-
}
|
|
|
39 |
},
|
40 |
{
|
41 |
"cell_type": "markdown",
|
|
|
|
|
|
|
42 |
"metadata": {
|
43 |
"id": "hcfn1TDmVqqg"
|
44 |
+
},
|
45 |
+
"source": [
|
46 |
+
"Let's also login to the Hugging Face Hub to have access to the Inference API."
|
47 |
+
]
|
48 |
},
|
49 |
{
|
50 |
"cell_type": "code",
|
|
|
725 |
},
|
726 |
{
|
727 |
"cell_type": "markdown",
|
728 |
+
"metadata": {
|
729 |
+
"id": "waE-prPUVzrh"
|
730 |
+
},
|
731 |
"source": [
|
732 |
"When you run this example, the output will **display a trace of the workflow steps being executed**. It will also print the corresponding Python code with the message:\n",
|
733 |
"\n",
|
|
|
739 |
"```\n",
|
740 |
"\n",
|
741 |
"After a few steps, you'll see the generated playlist that Alfred can use for the party! 🎵"
|
742 |
+
]
|
|
|
|
|
|
|
743 |
},
|
744 |
{
|
745 |
"cell_type": "markdown",
|
|
|
1021 |
" \"\"\"\n",
|
1022 |
" Suggests a menu based on the occasion.\n",
|
1023 |
" Args:\n",
|
1024 |
+
" occasion: The type of occasion for the party. enums: ['casual', 'formal', 'superhero', 'others']\n",
|
1025 |
" \"\"\"\n",
|
1026 |
" if occasion == \"casual\":\n",
|
1027 |
" return \"Pizza, snacks, and drinks.\"\n",
|
|
|
1039 |
},
|
1040 |
{
|
1041 |
"cell_type": "markdown",
|
1042 |
+
"metadata": {
|
1043 |
+
"id": "QZmuYtGPWKfO"
|
1044 |
+
},
|
1045 |
"source": [
|
1046 |
"The agent will run for a few steps until finding the answer.\n",
|
1047 |
"\n",
|
1048 |
"The menu is ready! 🥗"
|
1049 |
+
]
|
|
|
|
|
|
|
1050 |
},
|
1051 |
{
|
1052 |
"cell_type": "markdown",
|
|
|
1590 |
},
|
1591 |
{
|
1592 |
"cell_type": "markdown",
|
1593 |
+
"metadata": {
|
1594 |
+
"id": "T7pLiWCcWSsU"
|
1595 |
+
},
|
1596 |
"source": [
|
1597 |
"These examples are just the beginning of what you can do with code agents, and we're already starting to see their utility for preparing the party.\n",
|
1598 |
"You can learn more about how to build code agents in the [smolagents documentation](https://huggingface.co/docs/smolagents).\n",
|
1599 |
"\n",
|
1600 |
"`smolagents` specializes in agents that write and execute Python code snippets, offering sandboxed execution for security. It supports both open-source and proprietary language models, making it adaptable to various development environments."
|
1601 |
+
]
|
|
|
|
|
|
|
1602 |
},
|
1603 |
{
|
1604 |
"cell_type": "markdown",
|
|
|
3555 |
},
|
3556 |
{
|
3557 |
"cell_type": "markdown",
|
|
|
|
|
|
|
3558 |
"metadata": {
|
3559 |
"id": "3iPLK1IQWguH"
|
3560 |
+
},
|
3561 |
+
"source": [
|
3562 |
+
"To download the agent again, use the code below:"
|
3563 |
+
]
|
3564 |
},
|
3565 |
{
|
3566 |
"cell_type": "code",
|
|
|
4539 |
},
|
4540 |
{
|
4541 |
"cell_type": "markdown",
|
4542 |
+
"metadata": {
|
4543 |
+
"id": "qz-vwxu9Wobn"
|
4544 |
+
},
|
4545 |
"source": [
|
4546 |
"What's also exciting is that shared agents are directly available as Hugging Face Spaces, allowing you to interact with them in real-time. You can explore other agents [here](https://huggingface.co/spaces/davidberenstein1957/smolagents-and-tools).\n",
|
4547 |
"\n",
|
4548 |
"For example, the _AlfredAgent_ is available [here](https://huggingface.co/spaces/sergiopaniego/AlfredAgent)."
|
4549 |
+
]
|
|
|
|
|
|
|
4550 |
},
|
4551 |
{
|
4552 |
"cell_type": "markdown",
|
|
|
4580 |
},
|
4581 |
{
|
4582 |
"cell_type": "markdown",
|
4583 |
+
"metadata": {
|
4584 |
+
"id": "IVBwrNgbXPS-"
|
4585 |
+
},
|
4586 |
"source": [
|
4587 |
"Next, Alfred has already created an account on Langfuse and has his API keys ready. If you haven’t done so yet, you can sign up for Langfuse Cloud [here](https://cloud.langfuse.com/) or explore [alternatives](https://huggingface.co/docs/smolagents/tutorials/inspect_runs). \n",
|
4588 |
"\n",
|
4589 |
"Once you have your API keys, they need to be properly configured as follows:"
|
4590 |
+
]
|
|
|
|
|
|
|
4591 |
},
|
4592 |
{
|
4593 |
"cell_type": "code",
|
|
|
4612 |
},
|
4613 |
{
|
4614 |
"cell_type": "markdown",
|
|
|
|
|
|
|
4615 |
"metadata": {
|
4616 |
"id": "ZIRRj-PJXRGH"
|
4617 |
+
},
|
4618 |
+
"source": [
|
4619 |
+
"Finally, Alfred is ready to initialize the `SmolagentsInstrumentor` and start tracking his agent's performance. "
|
4620 |
+
]
|
4621 |
},
|
4622 |
{
|
4623 |
"cell_type": "code",
|
|
|
4641 |
},
|
4642 |
{
|
4643 |
"cell_type": "markdown",
|
|
|
|
|
|
|
4644 |
"metadata": {
|
4645 |
"id": "B0oLBr_qXSjA"
|
4646 |
+
},
|
4647 |
+
"source": [
|
4648 |
+
"Alfred is now connected 🔌! The runs from `smolagents` are being logged in Langfuse, giving him full visibility into the agent's behavior. With this setup, he's ready to revisit previous runs and refine his Party Preparator Agent even further. "
|
4649 |
+
]
|
4650 |
},
|
4651 |
{
|
4652 |
"cell_type": "code",
|
|
|
5212 |
},
|
5213 |
{
|
5214 |
"cell_type": "markdown",
|
5215 |
+
"metadata": {
|
5216 |
+
"id": "KBpqTBK1XVLz"
|
5217 |
+
},
|
5218 |
"source": [
|
5219 |
"Alfred can now access this logs [here](https://cloud.langfuse.com/project/cm7bq0abj025rad078ak3luwi/traces/995fc019255528e4f48cf6770b0ce27b?timestamp=2025-02-19T10%3A28%3A36.929Z) to review and analyze them. \n",
|
5220 |
"\n",
|
5221 |
"Meanwhile, the [suggested playlist](https://open.spotify.com/playlist/0gZMMHjuxMrrybQ7wTMTpw) sets the perfect vibe for the party preparations. Cool, right? 🎶\n"
|
5222 |
+
]
|
|
|
|
|
|
|
5223 |
}
|
5224 |
],
|
5225 |
"metadata": {
|
|
|
10631 |
},
|
10632 |
"nbformat": 4,
|
10633 |
"nbformat_minor": 0
|
10634 |
+
}
|