code snippets mismatch across different tasks

[not-lain]

(mostly due to inputs needing to be replaced by text in the python client, but it's not limited to that)
here are some examples of some of my findings

• text-classification
  this does not work
  
  as for the hub implementation

from huggingface_hub import InferenceClient

client = InferenceClient(
	provider="hf-inference",
+	model="ProsusAI/finbert",
+	provider="hf-inference",
	api_key="hf_xxxxxxxxxxxxxxxxxxxxxxxx"
)

result = client.text_classification(
-	model="ProsusAI/finbert",
-	inputs="I like you. I love you",
- 	provider="hf-inference",
+	text="I like you. I love you",
)

print(result)

• translation

• python

from huggingface_hub import InferenceClient

client = InferenceClient(
+	model="facebook/mbart-large-50-many-to-many-mmt",
	provider="hf-inference",
	api_key="hf_xxxxxxxxxxxxxxxxxxxxxxxx"
)

result = client.translation(
-	model="facebook/mbart-large-50-many-to-many-mmt",
-	inputs="Меня зовут Вольфганг и я живу в Берлине",
+ 	text ="Меня зовут Вольфганг и я живу в Берлине", 
+ 	src_lang="ru_RU",
+ 	tgt_lang ="en_XX"
-	provider="hf-inference",
)

print(result)

• js

import { HfInference } from "@huggingface/inference";

const client = new HfInference("hf_xxxxxxxxxxxxxxxxxxxxxxxx");

const output = await client.translation({
    model: "facebook/mbart-large-50-many-to-many-mmt",
    provider: "hf-inference",
    inputs: "comment allez vous ?",
+    parameters: {
+        src_lang: "fr_XX",
+        tgt_lang: "en_XX"
+    }
});

console.log(output);

[coyotte508]

cc @Wauplin for viz