try new param syntax w inference

README.md

@@ -12,6 +12,7 @@ hf_oauth_scopes:
   - inference-api
 models:
 # - gpt-3.5-turbo
+- bigscience/bloom-560m
 - Xenova/distilgpt2
 # - HuggingFaceH4/zephyr-7b-gemma-v0.1
 # - HuggingFaceH4/zephyr-7b-beta

sketch.js

@@ -4,7 +4,7 @@
 // import { pipeline, env } from 'https://cdn.jsdelivr.net/npm/@xenova/transformers';
 
 /// AUTHORIZATION
-
+import { textGeneration } from 'https://esm.sh/@huggingface/inference';
 import { oauthLoginUrl, oauthHandleRedirectIfPresent } from 'https://esm.sh/@huggingface/hub';
 
 const oauthResult = await oauthHandleRedirectIfPresent();
@@ -16,11 +16,11 @@ if (!oauthResult) {
 
 // You can use oauthResult.accessToken, oauthResult.accessTokenExpiresAt and oauthResult.userInfo
 // console.log(oauthResult);
-const HFAUTH = oauthResult.accessToken
-console.log(HFAUTH)
+const HF_TOKEN = oauthResult.accessToken
+console.log(HF_TOKEN)
 
-import { HfInference } from 'https://esm.sh/@huggingface/inference';
-const inference = new HfInference(HFAUTH);
+// import { HfInference } from 'https://esm.sh/@huggingface/inference';
+// const inference = new HfInference(HF_TOKEN);
 
 // PIPELINE MODELS
 // models('Xenova/gpt2', 'Xenova/gpt-3.5-turbo', 'mistralai/Mistral-7B-Instruct-v0.2', 'Xenova/llama-68m', 'meta-llama/Meta-Llama-3-8B', 'Xenova/bloom-560m', 'Xenova/distilgpt2')
@@ -179,17 +179,30 @@ async function runModel(PREPROMPT, PROMPT){
     // let MODELNAME = "HuggingFaceH4/zephyr-7b-beta"
     // let MODELNAME = "openai-community/gpt2"
     // let MODELNAME = 'mistral_inference'
-    let MODELNAME = 'Xenova/distilgpt2'
+    // let MODELNAME = 'Xenova/distilgpt2'
+    let MODELNAME = 'bigscience/bloom-560m'
 
-    let out = await inference.textGeneration(INPUT, {
+    let out = await textGeneration({    
+        accessToken: HF_TOKEN,
         model: MODELNAME,
-        max_new_tokens: 128
-    })
+        inputs: INPUT,
+        parameters: {
+            max_new_tokens: 128
+        }
+    });
+
+    // let out = await inference.textGeneration(INPUT, {
+    //     model: MODELNAME,
+    //     max_new_tokens: 128
+    // })
+
+    // let out = await inference.textGeneration(INPUT, 'bigscience/bloom-560m')
 
     // text-generation-inference
     // Uncaught (in promise) Error: HfApiJson(Deserialize(Error("unknown variant `transformers.js`, expected one of `text-generation-inference`, `transformers`, `allennlp`, `flair`, `espnet`, `asteroid`, `speechbrain`, `timm`, `sentence-transformers`, `spacy`, `sklearn`, `stanza`, `adapter-transformers`, `fasttext`, `fairseq`, `pyannote-audio`, `doctr`, `nemo`, `fastai`, `k2`, `diffusers`, `paddlenlp`, `mindspore`, `open_clip`, `span-marker`, `bertopic`, `peft`, `setfit`", line: 1, column: 397)))
 
     // let out = await inference.textGeneration({    
+    //     accessToken: HF_TOKEN,
     //     model: MODELNAME,
     //     messages: [{ 
     //                 role: "system", 
@@ -203,10 +216,12 @@ async function runModel(PREPROMPT, PROMPT){
 
     console.log(out)
 
+    console.log(out.token.text, out.generated_text)
+
     // modelResult = await out.messages[0].content
 
-    var modelResult = await out.choices[0].message.content
-    // var modelResult = await out[0].generated_text
+    // var modelResult = await out.choices[0].message.content
+    var modelResult = await out[0].generated_text
     console.log(modelResult);
 
     return modelResult
@@ -247,7 +262,7 @@ async function runModel(PREPROMPT, PROMPT){
 //     // let MODELNAME = 'mistralai/Mistral-Nemo-Instruct-2407'
 //     let MODELNAME = "HuggingFaceH4/zephyr-7b-beta"
 
-//     // HFAUTH
+//     // HF_TOKEN
 
 //     // 'meta-llama/Meta-Llama-3-70B-Instruct'
 //     // 'openai-community/gpt2'
@@ -259,7 +274,7 @@ async function runModel(PREPROMPT, PROMPT){
 //     // pipeline/transformers version
 //     let pipe = await pipeline('text-generation', {
 //         model: MODELNAME,
-//         accessToken: HFAUTH
+//         accessToken: HF_TOKEN
 //     });
 //     // seems to work with default model distilgpt2 ugh