[Inference snippets] Templated snippets for inference snippet generation #1255

Wauplin · 2025-03-06T17:54:11Z

What does this PR do?

add @huggingface/jinja as a dependency for @huggingface/inference
heavily refactor ./packages/inference/src/snippets/python.ts to use templates instead of inline strings
add some snippet templates under packages/inference/src/snippets/templates/{language}/{client}/{templateName}.jinja
- language is Python
- client is e.g. huggingface_hub, requests, etc.
- templateName is related to task name e.g. textToImage, conversationalStream, etc.

=> overall goal is to make it rather easy to update the snippets for a given task + client without entering the JS code
=> if conclusive, we can extend to cURL / JS snippets in a follow-up PR

This PR also includes a lot of small fixes (usually whitespace<>tabs inconsistencies) + added some tests. I did not change anything major in the generated snippets.

What to review?

IMO, no need to review all templates. Better to check the generated snippets + the main JS file directly (https://github.com/huggingface/huggingface.js/blob/ddd62b60d0a858dc808b6dd7b3589977681f5a8b/packages/inference/src/snippets/python.ts). Sorry for the huge PR but I do believe it is a necessary move moving forward 🤗

julien-c · 2025-03-06T18:18:56Z

packages/inference/package.json

@@ -52,9 +52,11 @@
 		"check": "tsc"
 	},
 	"dependencies": {
-		"@huggingface/tasks": "workspace:^"
+		"@huggingface/tasks": "workspace:^",
+		"handlebars": "^4.7.8"


or we could use https://www.npmjs.com/package/@huggingface/jinja maybe 🤯

That would be awesome! 🤩 Although the library was originally designed for ChatML templates, the set of available features should be large enough for these templates.

Maybe @Wauplin can explain what set of features would be required? 👀

basically just ifs and variable replacement from what i've seen

handlebars has pretty much a feature set of 0.00

Will update my PR tomorrow in that direction. As Julien said, I'm not using anything fancy at all so jinja will be more than enough for the job

Sounds good! Let me know if I can help in any way 🫡

I replaced the handlebars dependency by huggingface/jinja and it works like a charm! 86e787a Thanks for the package @xenova! 🤗

well I now have error Couldn't find package "@huggingface/jinja@^0.3.3" required by "@huggingface/inference@*" on the "npm" registry. in the CI though jinja 0.3.3 is available on https://www.npmjs.com/package/@huggingface/jinja 🤔

you need to add jinja here:

huggingface.js/.github/workflows/test.yml

Line 108 in 722b902

pnpm --filter inference --filter hub --filter tasks publish --force --no-git-checks --registry http://localhost:4874/

.vscode/settings.json

packages/inference/src/snippets/python.ts

packages/inference/src/snippets/templates/python/requests/textToAudio.jinja

Co-authored-by: Simon Brandeis <[email protected]>

packages/inference/src/lib/makeRequestOptions.ts

packages/inference/src/snippets/templates/python/fal_client/textToImage.jinja

…ingface/huggingface.js into templated-inference-python-snippets

coyotte508 · 2025-03-07T14:47:31Z

packages/inference/src/snippets/python.ts

+// Helpers to find + load templates
+
+const rootDirFinder = (): string => {
+	let currentPath = path.normalize(import.meta.url).replace("file:", "");


I'm curious if this will work with moon (import.meta.url), or if it's finally time to move from CJS to ESM cc @Pierrci

are you simply curious or actually wishing for it? x)

(I think it will be fine, but I may be wrong!)

I think it'll be empty but we'll see!

would f6d81d6 be ok ?

@coyotte508

⚠️ PR opened on top of #1255 Related to this thread #1255 (comment). @coyotte508 @SBrandeis @julien-c WDYT? 🙈 for the record, I'm looking for a solution where: 1. we keep this public/open-source 2. we keep the structure of jinja templates 3. ideally no new package in hf.js 4. ideally no new tooling (e.g. to translate jinja into TS code) ~Solution here is simply to gracefully raise an error if environment not supported.~ **EDIT:** based on #1259 (comment), goal is to not package the `./snippets` module in browser mode. --------- Co-authored-by: coyotte508 <[email protected]>

Wauplin · 2025-03-11T11:42:48Z

@coyotte508 could you check the last commits I've pushed to fix the CI? CI's now green but not entirely sure if I fixed things the correct way.

Otherwise @julien-c @hanouticelina @SBrandeis would you have time for a review of the snippets themselves? As mentioned in the PR description I think it's best to review the generated snippets under packages/tasks-gen/snippets-fixtures/ rather than the jinja templates.

hanouticelina

reviewed the generated snippets, all good 👍

Wauplin · 2025-03-12T09:24:15Z

I'm closing this PR as the structure of it is approved and I need it for a follow-up PR. If anyone spots an error in the templates, just comment it here and I'll take care in another PR.

@coyotte508

This PR should fix huggingface-internal/moon-landing#13013. It partially removes the structure introduced in #1255. Current problem is that inference snippets are generated in the front-end. Since front-end cannot access file-system and therefore the jinja files, we have to find a workaround. This PR adds a build step which exports all jinja files into a single TS module. I've updated the `package.json` file so that now the snippets code should be available in any environment (both node and browser). cc @coyotte508 who suggested such a solution. Tested it in `@tasks-gen` and "it works" --- For the record, the exported file (not committed in this PR) looks like this: _# packages/inference/src/snippets/templates.exported.ts_ ```ts // Generated file - do not edit directly export const templates: Record<string, Record<string, Record<string, string>>> = { "js": { "fetch": { "basic": "async function query(data) {\n\tconst response = await fetch(\n\t\t\"{{ fullUrl }}\",\n\t\t{\n\t\t\theaders: {\n\t\t\t\tAuthorization: \"{{ authorizationHeader }}\",\n\t\t\t\t\"Content-Type\": \"application/json\",\n\t\t\t},\n\t\t\tmethod: \"POST\",\n\t\t\tbody: JSON.stringify(data),\n\t\t}\n\t);\n\tconst result = await response.json();\n\treturn result;\n}\n\nquery({ inputs: {{ providerInputs.asObj.inputs }} }).then((response) => {\n console.log(JSON.stringify(response));\n});", "basicAudio": "async function query(data) {\n\tconst response = await fetch(\n\t\t\"{{ fullUrl }}\",\n\t\t{\n\t\t\theaders: {\n\t\t\t\tAuthorization: \"{{ authorizationHeader }}\",\n\t\t\t\t\"Content-Type\": \"audio/flac\"\n\t\t\t},\n\t\t\tmethod: \"POST\",\n\t\t\tbody: JSON.stringify(data),\n\t\t}\n\t);\n\tconst result = await response.json();\n\treturn result;\n}\n\nquery({ inputs: {{ providerInputs.asObj.inputs }} }).then((response) => {\n console.log(JSON.stringify(response));\n});", "basicImage": "async function query(data) {\n\tconst response = await fetch(\n\t\t\"{{ fullUrl }}\",\n\t\t{\n\t\t\theaders: {\n\t\t\t\tAuthorization: \"{{ authorizationHeader }}\",\n\t\t\t\t\"Content-Type\": \"image/jpeg\"\n\t\t\t},\n\t\t\tmethod: \"POST\",\n\t\t\tbody: JSON.stringify(data),\n\t\t}\n\t);\n\tconst result = await response.json();\n\treturn result;\n}\n\nquery({ inputs: {{ providerInputs.asObj.inputs }} }).then((response) => {\n console.log(JSON.stringify(response));\n});", "textToAudio": "{% if model.library_name == \"transformers\" %}\nasync function query(data) {\n\tconst response = await fetch(\n\t\t\"{{ fullUrl }}\",\n\t\t{\n\t\t\theaders: {\n\t\t\t\tAuthorization: \"{{ authorizationHeader }}\",\n\t\t\t\t\"Content-Type\": \"application/json\",\n\t\t\t},\n\t\t\tmethod: \"POST\",\n\t\t\tbody: JSON.stringify(data),\n\t\t}\n\t);\n\tconst result = await response.blob();\n return result;\n}\n\nquery({ inputs: {{ providerInputs.asObj.inputs }} }).then((response) => {\n // Returns a byte object of the Audio wavform. Use it directly!\n});\n{% else %}\nasync function query(data) {\n\tconst response = await fetch(\n\t\t\"{{ fullUrl }}\",\n\t\t{\n\t\t\theaders: {\n\t\t\t\tAuthorization: \"{{ authorizationHeader }}\",\n\t\t\t\t\"Content-Type\": \"application/json\",\n\t\t\t},\n\t\t\tmethod: \"POST\",\n\t\t\tbody: JSON.stringify(data),\n\t\t}\n\t);\n const result = await response.json();\n return result;\n}\n\nquery({ inputs: {{ providerInputs.asObj.inputs }} }).then((response) => {\n console.log(JSON.stringify(response));\n});\n{% endif %} ", "textToImage": "async function query(data) {\n\tconst response = await fetch(\n\t\t\"{{ fullUrl }}\",\n\t\t{\n\t\t\theaders: {\n\t\t\t\tAuthorization: \"{{ authorizationHeader }}\",\n\t\t\t\t\"Content-Type\": \"application/json\",\n\t\t\t},\n\t\t\tmethod: \"POST\",\n\t\t\tbody: JSON.stringify(data),\n\t\t}\n\t);\n\tconst result = await response.blob();\n\treturn result;\n}\n\nquery({ inputs: {{ providerInputs.asObj.inputs }} }).then((response) => {\n // Use image\n});", "zeroShotClassification": "async function query(data) {\n const response = await fetch(\n\t\t\"{{ fullUrl }}\",\n {\n headers: {\n\t\t\t\tAuthorization: \"{{ authorizationHeader }}\",\n \"Content-Type\": \"application/json\",\n },\n method: \"POST\",\n body: JSON.stringify(data),\n }\n );\n const result = await response.json();\n return result;\n}\n\nquery({\n inputs: {{ providerInputs.asObj.inputs }},\n parameters: { candidate_labels: [\"refund\", \"legal\", \"faq\"] }\n}).then((response) => {\n console.log(JSON.stringify(response));\n});" }, "huggingface.js": { "basic": "import { InferenceClient } from \"@huggingface/inference\";\n\nconst client = new InferenceClient(\"{{ accessToken }}\");\n\nconst output = await client.{{ methodName }}({\n\tmodel: \"{{ model.id }}\",\n\tinputs: {{ inputs.asObj.inputs }},\n\tprovider: \"{{ provider }}\",\n});\n\nconsole.log(output);", "basicAudio": "import { InferenceClient } from \"@huggingface/inference\";\n\nconst client = new InferenceClient(\"{{ accessToken }}\");\n\nconst data = fs.readFileSync({{inputs.asObj.inputs}});\n\nconst output = await client.{{ methodName }}({\n\tdata,\n\tmodel: \"{{ model.id }}\",\n\tprovider: \"{{ provider }}\",\n});\n\nconsole.log(output);", "basicImage": "import { InferenceClient } from \"@huggingface/inference\";\n\nconst client = new InferenceClient(\"{{ accessToken }}\");\n\nconst data = fs.readFileSync({{inputs.asObj.inputs}});\n\nconst output = await client.{{ methodName }}({\n\tdata,\n\tmodel: \"{{ model.id }}\",\n\tprovider: \"{{ provider }}\",\n});\n\nconsole.log(output);", "conversational": "import { InferenceClient } from \"@huggingface/inference\";\n\nconst client = new InferenceClient(\"{{ accessToken }}\");\n\nconst chatCompletion = await client.chatCompletion({\n provider: \"{{ provider }}\",\n model: \"{{ model.id }}\",\n{{ inputs.asTsString }}\n});\n\nconsole.log(chatCompletion.choices[0].message);", "conversationalStream": "import { InferenceClient } from \"@huggingface/inference\";\n\nconst client = new InferenceClient(\"{{ accessToken }}\");\n\nlet out = \"\";\n\nconst stream = await client.chatCompletionStream({\n provider: \"{{ provider }}\",\n model: \"{{ model.id }}\",\n{{ inputs.asTsString }}\n});\n\nfor await (const chunk of stream) {\n\tif (chunk.choices && chunk.choices.length > 0) {\n\t\tconst newContent = chunk.choices[0].delta.content;\n\t\tout += newContent;\n\t\tconsole.log(newContent);\n\t} \n}", "textToImage": "import { InferenceClient } from \"@huggingface/inference\";\n\nconst client = new InferenceClient(\"{{ accessToken }}\");\n\nconst image = await client.textToImage({\n provider: \"{{ provider }}\",\n model: \"{{ model.id }}\",\n\tinputs: {{ inputs.asObj.inputs }},\n\tparameters: { num_inference_steps: 5 },\n});\n/// Use the generated image (it's a Blob)", "textToVideo": "import { InferenceClient } from \"@huggingface/inference\";\n\nconst client = new InferenceClient(\"{{ accessToken }}\");\n\nconst image = await client.textToVideo({\n provider: \"{{ provider }}\",\n model: \"{{ model.id }}\",\n\tinputs: {{ inputs.asObj.inputs }},\n});\n// Use the generated video (it's a Blob)" }, "openai": { "conversational": "import { OpenAI } from \"openai\";\n\nconst client = new OpenAI({\n\tbaseURL: \"{{ baseUrl }}\",\n\tapiKey: \"{{ accessToken }}\",\n});\n\nconst chatCompletion = await client.chat.completions.create({\n\tmodel: \"{{ providerModelId }}\",\n{{ inputs.asTsString }}\n});\n\nconsole.log(chatCompletion.choices[0].message);", "conversationalStream": "import { OpenAI } from \"openai\";\n\nconst client = new OpenAI({\n\tbaseURL: \"{{ baseUrl }}\",\n\tapiKey: \"{{ accessToken }}\",\n});\n\nlet out = \"\";\n\nconst stream = await client.chat.completions.create({\n provider: \"{{ provider }}\",\n model: \"{{ model.id }}\",\n{{ inputs.asTsString }}\n});\n\nfor await (const chunk of stream) {\n\tif (chunk.choices && chunk.choices.length > 0) {\n\t\tconst newContent = chunk.choices[0].delta.content;\n\t\tout += newContent;\n\t\tconsole.log(newContent);\n\t} \n}" } }, "python": { "fal_client": { "textToImage": "{% if provider == \"fal-ai\" %}\nimport fal_client\n\nresult = fal_client.subscribe(\n \"{{ providerModelId }}\",\n arguments={\n \"prompt\": {{ inputs.asObj.inputs }},\n },\n)\nprint(result)\n{% endif %} " }, "huggingface_hub": { "basic": "result = client.{{ methodName }}(\n inputs={{ inputs.asObj.inputs }},\n model=\"{{ model.id }}\",\n)", "basicAudio": "output = client.{{ methodName }}({{ inputs.asObj.inputs }}, model=\"{{ model.id }}\")", "basicImage": "output = client.{{ methodName }}({{ inputs.asObj.inputs }}, model=\"{{ model.id }}\")", "conversational": "completion = client.chat.completions.create(\n model=\"{{ model.id }}\",\n{{ inputs.asPythonString }}\n)\n\nprint(completion.choices[0].message) ", "conversationalStream": "stream = client.chat.completions.create(\n model=\"{{ model.id }}\",\n{{ inputs.asPythonString }}\n stream=True,\n)\n\nfor chunk in stream:\n print(chunk.choices[0].delta.content, end=\"\") ", "documentQuestionAnswering": "output = client.document_question_answering(\n \"{{ inputs.asObj.image }}\",\n question=\"{{ inputs.asObj.question }}\",\n model=\"{{ model.id }}\",\n) ", "imageToImage": "# output is a PIL.Image object\nimage = client.image_to_image(\n \"{{ inputs.asObj.inputs }}\",\n prompt=\"{{ inputs.asObj.parameters.prompt }}\",\n model=\"{{ model.id }}\",\n) ", "importInferenceClient": "from huggingface_hub import InferenceClient\n\nclient = InferenceClient(\n provider=\"{{ provider }}\",\n api_key=\"{{ accessToken }}\",\n)", "textToImage": "# output is a PIL.Image object\nimage = client.text_to_image(\n {{ inputs.asObj.inputs }},\n model=\"{{ model.id }}\",\n) ", "textToVideo": "video = client.text_to_video(\n {{ inputs.asObj.inputs }},\n model=\"{{ model.id }}\",\n) " }, "openai": { "conversational": "from openai import OpenAI\n\nclient = OpenAI(\n base_url=\"{{ baseUrl }}\",\n api_key=\"{{ accessToken }}\"\n)\n\ncompletion = client.chat.completions.create(\n model=\"{{ providerModelId }}\",\n{{ inputs.asPythonString }}\n)\n\nprint(completion.choices[0].message) ", "conversationalStream": "from openai import OpenAI\n\nclient = OpenAI(\n base_url=\"{{ baseUrl }}\",\n api_key=\"{{ accessToken }}\"\n)\n\nstream = client.chat.completions.create(\n model=\"{{ providerModelId }}\",\n{{ inputs.asPythonString }}\n stream=True,\n)\n\nfor chunk in stream:\n print(chunk.choices[0].delta.content, end=\"\")" }, "requests": { "basic": "def query(payload):\n response = requests.post(API_URL, headers=headers, json=payload)\n return response.json()\n\noutput = query({\n \"inputs\": {{ providerInputs.asObj.inputs }},\n}) ", "basicAudio": "def query(filename):\n with open(filename, \"rb\") as f:\n data = f.read()\n response = requests.post(API_URL, headers={\"Content-Type\": \"audio/flac\", **headers}, data=data)\n return response.json()\n\noutput = query({{ providerInputs.asObj.inputs }})", "basicImage": "def query(filename):\n with open(filename, \"rb\") as f:\n data = f.read()\n response = requests.post(API_URL, headers={\"Content-Type\": \"image/jpeg\", **headers}, data=data)\n return response.json()\n\noutput = query({{ providerInputs.asObj.inputs }})", "conversational": "def query(payload):\n response = requests.post(API_URL, headers=headers, json=payload)\n return response.json()\n\nresponse = query({\n{{ providerInputs.asJsonString }}\n})\n\nprint(response[\"choices\"][0][\"message\"])", "conversationalStream": "def query(payload):\n response = requests.post(API_URL, headers=headers, json=payload, stream=True)\n for line in response.iter_lines():\n if not line.startswith(b\"data:\"):\n continue\n if line.strip() == b\"data: [DONE]\":\n return\n yield json.loads(line.decode(\"utf-8\").lstrip(\"data:\").rstrip(\"/n\"))\n\nchunks = query({\n{{ providerInputs.asJsonString }},\n \"stream\": True,\n})\n\nfor chunk in chunks:\n print(chunk[\"choices\"][0][\"delta\"][\"content\"], end=\"\")", "documentQuestionAnswering": "def query(payload):\n with open(payload[\"image\"], \"rb\") as f:\n img = f.read()\n payload[\"image\"] = base64.b64encode(img).decode(\"utf-8\")\n response = requests.post(API_URL, headers=headers, json=payload)\n return response.json()\n\noutput = query({\n \"inputs\": {\n \"image\": \"{{ inputs.asObj.image }}\",\n \"question\": \"{{ inputs.asObj.question }}\",\n },\n}) ", "imageToImage": "def query(payload):\n with open(payload[\"inputs\"], \"rb\") as f:\n img = f.read()\n payload[\"inputs\"] = base64.b64encode(img).decode(\"utf-8\")\n response = requests.post(API_URL, headers=headers, json=payload)\n return response.content\n\nimage_bytes = query({\n{{ providerInputs.asJsonString }}\n})\n\n# You can access the image with PIL.Image for example\nimport io\nfrom PIL import Image\nimage = Image.open(io.BytesIO(image_bytes)) ", "importRequests": "{% if importBase64 %}\nimport base64\n{% endif %}\n{% if importJson %}\nimport json\n{% endif %}\nimport requests\n\nAPI_URL = \"{{ fullUrl }}\"\nheaders = {\"Authorization\": \"{{ authorizationHeader }}\"}", "tabular": "def query(payload):\n response = requests.post(API_URL, headers=headers, json=payload)\n return response.content\n\nresponse = query({\n \"inputs\": {\n \"data\": {{ providerInputs.asObj.inputs }}\n },\n}) ", "textToAudio": "{% if model.library_name == \"transformers\" %}\ndef query(payload):\n response = requests.post(API_URL, headers=headers, json=payload)\n return response.content\n\naudio_bytes = query({\n \"inputs\": {{ providerInputs.asObj.inputs }},\n})\n# You can access the audio with IPython.display for example\nfrom IPython.display import Audio\nAudio(audio_bytes)\n{% else %}\ndef query(payload):\n response = requests.post(API_URL, headers=headers, json=payload)\n return response.json()\n\naudio, sampling_rate = query({\n \"inputs\": {{ providerInputs.asObj.inputs }},\n})\n# You can access the audio with IPython.display for example\nfrom IPython.display import Audio\nAudio(audio, rate=sampling_rate)\n{% endif %} ", "textToImage": "{% if provider == \"hf-inference\" %}\ndef query(payload):\n response = requests.post(API_URL, headers=headers, json=payload)\n return response.content\n\nimage_bytes = query({\n \"inputs\": {{ providerInputs.asObj.inputs }},\n})\n\n# You can access the image with PIL.Image for example\nimport io\nfrom PIL import Image\nimage = Image.open(io.BytesIO(image_bytes))\n{% endif %}", "zeroShotClassification": "def query(payload):\n response = requests.post(API_URL, headers=headers, json=payload)\n return response.json()\n\noutput = query({\n \"inputs\": {{ providerInputs.asObj.inputs }},\n \"parameters\": {\"candidate_labels\": [\"refund\", \"legal\", \"faq\"]},\n}) ", "zeroShotImageClassification": "def query(data):\n with open(data[\"image_path\"], \"rb\") as f:\n img = f.read()\n payload={\n \"parameters\": data[\"parameters\"],\n \"inputs\": base64.b64encode(img).decode(\"utf-8\")\n }\n response = requests.post(API_URL, headers=headers, json=payload)\n return response.json()\n\noutput = query({\n \"image_path\": {{ providerInputs.asObj.inputs }},\n \"parameters\": {\"candidate_labels\": [\"cat\", \"dog\", \"llama\"]},\n}) " } }, "sh": { "curl": { "basic": "curl {{ fullUrl }} \\\n -X POST \\\n -H 'Authorization: {{ authorizationHeader }}' \\\n -H 'Content-Type: application/json' \\\n -d '{\n{{ providerInputs.asCurlString }}\n }'", "basicAudio": "curl {{ fullUrl }} \\\n -X POST \\\n -H 'Authorization: {{ authorizationHeader }}' \\\n -H 'Content-Type: audio/flac' \\\n --data-binary @{{ providerInputs.asObj.inputs }}", "basicImage": "curl {{ fullUrl }} \\\n -X POST \\\n -H 'Authorization: {{ authorizationHeader }}' \\\n -H 'Content-Type: image/jpeg' \\\n --data-binary @{{ providerInputs.asObj.inputs }}", "conversational": "curl {{ fullUrl }} \\\n -H 'Authorization: {{ authorizationHeader }}' \\\n -H 'Content-Type: application/json' \\\n -d '{\n{{ providerInputs.asCurlString }},\n \"stream\": false\n }'", "conversationalStream": "curl {{ fullUrl }} \\\n -H 'Authorization: {{ authorizationHeader }}' \\\n -H 'Content-Type: application/json' \\\n -d '{\n{{ providerInputs.asCurlString }},\n \"stream\": true\n }'", "zeroShotClassification": "curl {{ fullUrl }} \\\n -X POST \\\n -d '{\"inputs\": {{ providerInputs.asObj.inputs }}, \"parameters\": {\"candidate_labels\": [\"refund\", \"legal\", \"faq\"]}}' \\\n -H 'Content-Type: application/json' \\\n -H 'Authorization: {{ authorizationHeader }}'" } } } as const; ```

Wauplin added 19 commits March 5, 2025 17:29

tmp commit

ea68541

nope

0006042

draft

8ae6d93

fixing some tab -> whitespaces issues

b86faf5

first move towards templating

f121009

always more

c2231ad

all python snippets as templates + some examples

7090adc

base url for openai snippets

134fa2d

start refacto

d07032e

closer to a good version

2b95376

minimal code + trim all snippets

0766e08

some comments

29ebae9

sort

95a2a71

sort

f529ba6

delete before coming back

b7b6869

moving generated snippets

f8ecf33

remove tmp script

ddd62b6

tabs -> whitespaces

7a498c6

tabs -> whitespaces

e23a2a3

Wauplin requested review from julien-c, hanouticelina, SBrandeis and coyotte508 as code owners March 6, 2025 17:54

Merge branch 'main' into templated-inference-python-snippets

e9a2c13

julien-c reviewed Mar 6, 2025

View reviewed changes

switch from handlebars to jinja

86e787a

SBrandeis reviewed Mar 7, 2025

View reviewed changes

.vscode/settings.json Outdated Show resolved Hide resolved

packages/inference/src/snippets/python.ts Show resolved Hide resolved

packages/inference/src/snippets/templates/python/requests/textToAudio.jinja Show resolved Hide resolved

Update .vscode/settings.json

30bbfad

Co-authored-by: Simon Brandeis <[email protected]>

julien-c reviewed Mar 7, 2025

View reviewed changes

packages/inference/src/lib/makeRequestOptions.ts Outdated Show resolved Hide resolved

julien-c reviewed Mar 7, 2025

View reviewed changes

packages/inference/src/snippets/templates/python/fal_client/textToImage.jinja Show resolved Hide resolved

Wauplin and others added 3 commits March 7, 2025 13:04

not private

a280d60

Merge branch 'templated-inference-python-snippets' of github.com:hugg…

274dfe2

…ingface/huggingface.js into templated-inference-python-snippets

Merge branch 'main' into templated-inference-python-snippets

acb99d1

Wauplin mentioned this pull request Mar 7, 2025

Do not export ./snippets module in browser mode #1259

Merged

coyotte508 reviewed Mar 7, 2025

View reviewed changes

Wauplin and others added 6 commits March 10, 2025 18:45

Merge branch 'main' into templated-inference-python-snippets

99e8c87

Merge branch 'main' into templated-inference-python-snippets

dd96c31

filter jinja in ci test

aeecd11

handle both ESM and CJS ?

f6d81d6

add --allow-read flag for deno import

e4bacdd

coyotte508 approved these changes Mar 11, 2025

View reviewed changes

hanouticelina approved these changes Mar 11, 2025

View reviewed changes

Wauplin merged commit 2a5443f into main Mar 12, 2025
5 checks passed

Wauplin deleted the templated-inference-python-snippets branch March 12, 2025 09:24

Wauplin mentioned this pull request Mar 19, 2025

Export jinja files as TS module at build time #1296

Merged

[Inference snippets] Templated snippets for inference snippet generation #1255

[Inference snippets] Templated snippets for inference snippet generation #1255

Uh oh!

Conversation

Wauplin commented Mar 6, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

What does this PR do?

What to review?

Uh oh!

julien-c Mar 6, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Wauplin commented Mar 11, 2025

Uh oh!

hanouticelina left a comment

Choose a reason for hiding this comment

Uh oh!

Wauplin commented Mar 12, 2025

Uh oh!

Uh oh!

Uh oh!

Wauplin commented Mar 6, 2025 •

edited

Loading

julien-c Mar 6, 2025 •

edited

Loading