create

README.md

@@ -0,0 +1,57 @@
+## bloom-7b-rl
+
+This is a reinforcement learning enhanced bloom model (bloom-rl), fine-tuned based on bloom-7b (Muennighoff et al.).
+
+### Usage
+
+If you don't have a good GPU (mem > 20G) then use the code below:
+
+```python
+# pip install -q transformers accelerate
+from transformers import AutoModelForCausalLM, AutoTokenizer
+
+checkpoint = "hongyin/bloom-7b-rl"
+
+tokenizer = AutoTokenizer.from_pretrained(checkpoint)
+model = AutoModelForCausalLM.from_pretrained(checkpoint)
+
+inputs = tokenizer.encode("Translate to Chinese: I love you.", return_tensors="pt")
+outputs = model.generate(inputs)
+print(tokenizer.decode(outputs[0]))
+
+Translate to Chinese: I love you. 翻译:我爱你
+```
+
+If you have a good GPU (mem > 20G) then use the code below:
+
+```python
+# pip install -q transformers accelerate
+from transformers import AutoModelForCausalLM, AutoTokenizer
+
+checkpoint = "hongyin/bloom-7b-rl"
+
+tokenizer = AutoTokenizer.from_pretrained(checkpoint)
+model = AutoModelForCausalLM.from_pretrained(checkpoint, torch_dtype="auto", device_map="auto")
+
+inputs = tokenizer.encode("Translate to Chinese: I love you.", return_tensors="pt").to("cuda")
+outputs = model.generate(inputs)
+print(tokenizer.decode(outputs[0]))
+
+Translate to Chinese: I love you. 翻译:我爱你
+```
+
+## Bibtex entry and citation info
+Please cite if you find it helpful.
+```
+@article{zhu2023metaaid,
+  title={MetaAID 2.0: An Extensible Framework for Developing Metaverse Applications via Human-controllable Pre-trained Models},
+  author={Zhu, Hongyin},
+  journal={arXiv preprint arXiv:2302.13173},
+  year={2023}
+}
+
+```
+
+---
+license: other
+---

config.json

@@ -0,0 +1,34 @@
+{
+  "_name_or_path": "bloom-7b1-ckpt",
+  "apply_residual_connection_post_layernorm": false,
+  "architectures": [
+    "BloomForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "attention_softmax_in_fp32": true,
+  "bias_dropout_fusion": true,
+  "bos_token_id": 1,
+  "dropout": 0.0,
+  "end_token_id": 2,
+  "eos_token_id": 2,
+  "hidden_dropout": 0.0,
+  "hidden_size": 4096,
+  "initializer_range": 0.02,
+  "layer_norm_epsilon": 1e-05,
+  "masked_softmax_fusion": true,
+  "model_type": "bloom",
+  "n_head": 32,
+  "n_inner": null,
+  "n_layer": 30,
+  "offset_alibi": 100,
+  "pad_token_id": 2,
+  "pretraining_tp": 1,
+  "skip_bias_add": true,
+  "skip_bias_add_qkv": false,
+  "slow_but_exact": false,
+  "torch_dtype": "float32",
+  "transformers_version": "4.28.0",
+  "unk_token_id": 0,
+  "use_cache": true,
+  "vocab_size": 250680
+}

special_tokens_map.json

@@ -0,0 +1,6 @@
+{
+  "bos_token": "<s>",
+  "eos_token": "</s>",
+  "pad_token": "<pad>",
+  "unk_token": "<unk>"
+}

tokenizer_config.json

@@ -0,0 +1,11 @@
+{
+  "add_prefix_space": false,
+  "bos_token": "<s>",
+  "eos_token": "</s>",
+  "model_max_length": 1000000000000000019884624838656,
+  "pad_token": "<pad>",
+  "padding_side": "left",
+  "special_tokens_map_file": null,
+  "tokenizer_class": "BloomTokenizer",
+  "unk_token": "<unk>"
+}