Skip to content

Commit 135c53e

Browse files
Wauplinhanouticelina
Wauplin
and
hanouticelina
authored
Fix generate script after new snippets (#1490)
* Fix generate script after new snippets * [in progress] show all clients snippets * fix snippets * fix again * fix comment * separate snippets * separate snippets * refactor snippet fetching * (nit) improve client reference format --------- Co-authored-by: Celina Hanouti <[email protected]>
1 parent 23a1d86 commit 135c53e

File tree

8 files changed

+344
-113
lines changed

8 files changed

+344
-113
lines changed

‎docs/api-inference/tasks/chat-completion.md‎

Lines changed: 256 additions & 55 deletions
Original file line numberDiff line numberDiff line change
@@ -24,13 +24,12 @@ This is a subtask of [`text-generation`](https://huggingface.co/docs/api-inferen
2424
- [google/gemma-2-2b-it](https://huggingface.co/google/gemma-2-2b-it): A text-generation model trained to follow instructions.
2525
- [meta-llama/Meta-Llama-3.1-8B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3.1-8B-Instruct): Very powerful text generation model trained to follow instructions.
2626
- [microsoft/Phi-3-mini-4k-instruct](https://huggingface.co/microsoft/Phi-3-mini-4k-instruct): Small yet powerful text generation model.
27-
- [HuggingFaceH4/starchat2-15b-v0.1](https://huggingface.co/HuggingFaceH4/starchat2-15b-v0.1): Strong coding assistant model.
28-
- [mistralai/Mistral-Nemo-Instruct-2407](https://huggingface.co/mistralai/Mistral-Nemo-Instruct-2407): Very strong open-source large language model.
27+
- [Qwen/Qwen2.5-7B-Instruct](https://huggingface.co/Qwen/Qwen2.5-7B-Instruct): Strong text generation model to follow instructions.
2928

3029
#### Conversational Vision-Language Models (VLMs)
3130

3231
- [meta-llama/Llama-3.2-11B-Vision-Instruct](https://huggingface.co/meta-llama/Llama-3.2-11B-Vision-Instruct): Powerful vision language model with great visual understanding and reasoning capabilities.
33-
- [microsoft/Phi-3.5-vision-instruct](https://huggingface.co/microsoft/Phi-3.5-vision-instruct): Strong image-text-to-text model.
32+
- [Qwen/Qwen2-VL-7B-Instruct](https://huggingface.co/Qwen/Qwen2-VL-7B-Instruct): Strong image-text-to-text model.
3433

3534
### API Playground
3635

@@ -65,46 +64,133 @@ The API supports:
6564
curl 'https://api-inference.huggingface.co/models/google/gemma-2-2b-it/v1/chat/completions' \
6665
-H "Authorization: Bearer hf_***" \
6766
-H 'Content-Type: application/json' \
68-
-d '{
69-
"model": "google/gemma-2-2b-it",
70-
"messages": [{"role": "user", "content": "What is the capital of France?"}],
71-
"max_tokens": 500,
72-
"stream": false
67+
--data '{
68+
"model": "google/gemma-2-2b-it",
69+
"messages": [
70+
{
71+
"role": "user",
72+
"content": "What is the capital of France?"
73+
}
74+
],
75+
"max_tokens": 500,
76+
"stream": true
7377
}'
74-
7578
```
7679
</curl>
7780

7881
<python>
82+
With huggingface_hub client:
7983
```py
8084
from huggingface_hub import InferenceClient
8185

8286
client = InferenceClient(api_key="hf_***")
8387

84-
for message in client.chat_completion(
85-
model="google/gemma-2-2b-it",
86-
messages=[{"role": "user", "content": "What is the capital of France?"}],
88+
messages = [
89+
{
90+
"role": "user",
91+
"content": "What is the capital of France?"
92+
}
93+
]
94+
95+
stream = client.chat.completions.create(
96+
model="google/gemma-2-2b-it",
97+
messages=messages,
8798
max_tokens=500,
88-
stream=True,
89-
):
90-
print(message.choices[0].delta.content, end="")
99+
stream=True
100+
)
101+
102+
for chunk in stream:
103+
print(chunk.choices[0].delta.content, end="")
104+
```
105+
106+
With openai client:
107+
```py
108+
from openai import OpenAI
109+
110+
client = OpenAI(
111+
base_url="https://api-inference.huggingface.co/v1/",
112+
api_key="hf_***"
113+
)
114+
115+
messages = [
116+
{
117+
"role": "user",
118+
"content": "What is the capital of France?"
119+
}
120+
]
121+
122+
stream = client.chat.completions.create(
123+
model="google/gemma-2-2b-it",
124+
messages=messages,
125+
max_tokens=500,
126+
stream=True
127+
)
128+
129+
for chunk in stream:
130+
print(chunk.choices[0].delta.content, end="")
91131
```
92132

93133
To use the Python client, see `huggingface_hub`'s [package reference](https://huggingface.co/docs/huggingface_hub/package_reference/inference_client#huggingface_hub.InferenceClient.chat_completion).
94134
</python>
95135

96136
<js>
137+
With huggingface_hub client:
97138
```js
98-
import { HfInference } from "@huggingface/inference";
139+
import { HfInference } from "@huggingface/inference"
99140

100-
const inference = new HfInference("hf_***");
141+
const client = new HfInference("hf_***")
101142

102-
for await (const chunk of inference.chatCompletionStream({
143+
let out = "";
144+
145+
const stream = client.chatCompletionStream({
103146
model: "google/gemma-2-2b-it",
104-
messages: [{ role: "user", content: "What is the capital of France?" }],
147+
messages: [
148+
{
149+
role: "user",
150+
content: "What is the capital of France?"
151+
}
152+
],
153+
max_tokens: 500
154+
});
155+
156+
for await (const chunk of stream) {
157+
if (chunk.choices && chunk.choices.length > 0) {
158+
const newContent = chunk.choices[0].delta.content;
159+
out += newContent;
160+
console.log(newContent);
161+
}
162+
}
163+
```
164+
165+
With openai client:
166+
```js
167+
import { OpenAI } from "openai"
168+
169+
const client = new OpenAI({
170+
baseURL: "https://api-inference.huggingface.co/v1/",
171+
apiKey: "hf_***"
172+
})
173+
174+
let out = "";
175+
176+
const stream = await client.chat.completions.create({
177+
model: "google/gemma-2-2b-it",
178+
messages: [
179+
{
180+
role: "user",
181+
content: "What is the capital of France?"
182+
}
183+
],
105184
max_tokens: 500,
106-
})) {
107-
process.stdout.write(chunk.choices[0]?.delta?.content || "");
185+
stream: true,
186+
});
187+
188+
for await (const chunk of stream) {
189+
if (chunk.choices && chunk.choices.length > 0) {
190+
const newContent = chunk.choices[0].delta.content;
191+
out += newContent;
192+
console.log(newContent);
193+
}
108194
}
109195
```
110196

@@ -125,73 +211,188 @@ To use the JavaScript client, see `huggingface.js`'s [package reference](https:/
125211
curl 'https://api-inference.huggingface.co/models/meta-llama/Llama-3.2-11B-Vision-Instruct/v1/chat/completions' \
126212
-H "Authorization: Bearer hf_***" \
127213
-H 'Content-Type: application/json' \
128-
-d '{
129-
"model": "meta-llama/Llama-3.2-11B-Vision-Instruct",
130-
"messages": [
214+
--data '{
215+
"model": "meta-llama/Llama-3.2-11B-Vision-Instruct",
216+
"messages": [
131217
{
132218
"role": "user",
133219
"content": [
134-
{"type": "image_url", "image_url": {"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"}},
135-
{"type": "text", "text": "Describe this image in one sentence."}
220+
{
221+
"type": "text",
222+
"text": "Describe this image in one sentence."
223+
},
224+
{
225+
"type": "image_url",
226+
"image_url": {
227+
"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
228+
}
229+
}
136230
]
137231
}
138232
],
139-
"max_tokens": 500,
140-
"stream": false
233+
"max_tokens": 500,
234+
"stream": true
141235
}'
142-
143236
```
144237
</curl>
145238

146239
<python>
240+
With huggingface_hub client:
147241
```py
148242
from huggingface_hub import InferenceClient
149243

150244
client = InferenceClient(api_key="hf_***")
151245

152-
image_url = "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
246+
messages = [
247+
{
248+
"role": "user",
249+
"content": [
250+
{
251+
"type": "text",
252+
"text": "Describe this image in one sentence."
253+
},
254+
{
255+
"type": "image_url",
256+
"image_url": {
257+
"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
258+
}
259+
}
260+
]
261+
}
262+
]
263+
264+
stream = client.chat.completions.create(
265+
model="meta-llama/Llama-3.2-11B-Vision-Instruct",
266+
messages=messages,
267+
max_tokens=500,
268+
stream=True
269+
)
153270

154-
for message in client.chat_completion(
155-
model="meta-llama/Llama-3.2-11B-Vision-Instruct",
156-
messages=[
157-
{
158-
"role": "user",
159-
"content": [
160-
{"type": "image_url", "image_url": {"url": image_url}},
161-
{"type": "text", "text": "Describe this image in one sentence."},
162-
],
163-
}
164-
],
271+
for chunk in stream:
272+
print(chunk.choices[0].delta.content, end="")
273+
```
274+
275+
With openai client:
276+
```py
277+
from openai import OpenAI
278+
279+
client = OpenAI(
280+
base_url="https://api-inference.huggingface.co/v1/",
281+
api_key="hf_***"
282+
)
283+
284+
messages = [
285+
{
286+
"role": "user",
287+
"content": [
288+
{
289+
"type": "text",
290+
"text": "Describe this image in one sentence."
291+
},
292+
{
293+
"type": "image_url",
294+
"image_url": {
295+
"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
296+
}
297+
}
298+
]
299+
}
300+
]
301+
302+
stream = client.chat.completions.create(
303+
model="meta-llama/Llama-3.2-11B-Vision-Instruct",
304+
messages=messages,
165305
max_tokens=500,
166-
stream=True,
167-
):
168-
print(message.choices[0].delta.content, end="")
306+
stream=True
307+
)
308+
309+
for chunk in stream:
310+
print(chunk.choices[0].delta.content, end="")
169311
```
170312

171313
To use the Python client, see `huggingface_hub`'s [package reference](https://huggingface.co/docs/huggingface_hub/package_reference/inference_client#huggingface_hub.InferenceClient.chat_completion).
172314
</python>
173315

174316
<js>
317+
With huggingface_hub client:
175318
```js
176-
import { HfInference } from "@huggingface/inference";
319+
import { HfInference } from "@huggingface/inference"
177320

178-
const inference = new HfInference("hf_***");
179-
const imageUrl = "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg";
321+
const client = new HfInference("hf_***")
180322

181-
for await (const chunk of inference.chatCompletionStream({
323+
let out = "";
324+
325+
const stream = client.chatCompletionStream({
182326
model: "meta-llama/Llama-3.2-11B-Vision-Instruct",
183327
messages: [
184328
{
185-
"role": "user",
186-
"content": [
187-
{"type": "image_url", "image_url": {"url": imageUrl}},
188-
{"type": "text", "text": "Describe this image in one sentence."},
189-
],
329+
role: "user",
330+
content: [
331+
{
332+
type: "text",
333+
text: "Describe this image in one sentence."
334+
},
335+
{
336+
type: "image_url",
337+
image_url: {
338+
url: "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
339+
}
340+
}
341+
]
342+
}
343+
],
344+
max_tokens: 500
345+
});
346+
347+
for await (const chunk of stream) {
348+
if (chunk.choices && chunk.choices.length > 0) {
349+
const newContent = chunk.choices[0].delta.content;
350+
out += newContent;
351+
console.log(newContent);
352+
}
353+
}
354+
```
355+
356+
With openai client:
357+
```js
358+
import { OpenAI } from "openai"
359+
360+
const client = new OpenAI({
361+
baseURL: "https://api-inference.huggingface.co/v1/",
362+
apiKey: "hf_***"
363+
})
364+
365+
let out = "";
366+
367+
const stream = await client.chat.completions.create({
368+
model: "meta-llama/Llama-3.2-11B-Vision-Instruct",
369+
messages: [
370+
{
371+
role: "user",
372+
content: [
373+
{
374+
type: "text",
375+
text: "Describe this image in one sentence."
376+
},
377+
{
378+
type: "image_url",
379+
image_url: {
380+
url: "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
381+
}
382+
}
383+
]
190384
}
191385
],
192386
max_tokens: 500,
193-
})) {
194-
process.stdout.write(chunk.choices[0]?.delta?.content || "");
387+
stream: true,
388+
});
389+
390+
for await (const chunk of stream) {
391+
if (chunk.choices && chunk.choices.length > 0) {
392+
const newContent = chunk.choices[0].delta.content;
393+
out += newContent;
394+
console.log(newContent);
395+
}
195396
}
196397
```
197398

‎docs/api-inference/tasks/image-classification.md‎

Lines changed: 0 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -25,7 +25,6 @@ For more details about the `image-classification` task, check out its [dedicated
2525
### Recommended models
2626

2727
- [google/vit-base-patch16-224](https://huggingface.co/google/vit-base-patch16-224): A strong image classification model.
28-
- [facebook/deit-base-distilled-patch16-224](https://huggingface.co/facebook/deit-base-distilled-patch16-224): A robust image classification model.
2928

3029
Explore all available models and find the one that suits you best [here](https://huggingface.co/models?inference=warm&pipeline_tag=image-classification&sort=trending).
3130

0 commit comments

Comments
 (0)