Use TF model object as parameter in new redisai.Model

[CICCIOSGAMINO]

As a Node.js developer i'm using the redisai-js module to connect with Redis-Ai with my node.js application. So i'm using / save / load my models in TensorflowJs format (model.json, weights.bin) so its useful create the Redis-Ai model getting start from these two files saved. Actually the code i'm used doesn't work because the model format:

    const m = await tf.loadLayersModel(`file://model/AX-model/model.json`)
    const myModel = new redisai.Model(redisai.Backend.TF, 'CPU',['a','b'], ['c'], m)
    const r = await aiclient.modelset('mlmodel', myModel)

So i'm interesting to thinking at a solution, in node.js application i'm struggle to use Redis to save the TensorflowJs models too directly in-memory db. I'd like to chose which operation get done in Node.js application and which in Redis-Ai, so for example train the model in Node.js application and predict in Redis-Ai:

     const model = tf.sequential()
     model.add(tf.layers.dense({inputShape: [1], units: 1}))
     model.add(tf.layers.dense({units: 1}))
    model.compile({
           optimizer: 'sgd',   // tf.train.adam(),
           // loss: tf.losses.meanSquaredError,
           loss: 'meanSquaredError',
           metrics: ['mse']
      })

       const myModel = new redisai.Model(redisai.Backend.TF, 'CPU',['a','b'], ['c'], model)
       const r = await aiclient.modelset('mlmodel', myModel)

[chayim]

@leibale Want to have a look?