TypeError - Read only flag not supported - converttopytorch.py

[gbrackenridge]

I am struggling to get converttotorch.py running properly. I keep throwing a TypeError as show below. Looks like there was some noise in git over in the numpy code repo. Any ideas? Thanks in advance!

---

loading shards for part 0
read from checkpoint
< (8, 4096) to (4096,)

transformer.wte.bias torch.Size([4096])
< (8, 6300, 4096) to (1, 50400, 4096)
transformer.wte.weight torch.Size([4096, 50400])
< (8, 4096, 512) to (1, 4096, 4096)
Traceback (most recent call last):
File "/data/models/model38/gpt-j-6B/converttotorch.py", line 147, in
params = torch.tensor(params.copy()).half()
TypeError: the read only flag is not supported, should always be False

[NatLee]

Replace the line with this and it'll be fixed.

Before:

params = torch.tensor(params.copy()).half()

After:

params = torch.from_numpy(np.array(params)).half()