Upload 3 files

generate_hf.py

@@ -0,0 +1,72 @@
+import torch
+from peft import PeftModel
+import transformers
+import gradio as gr
+
+from ringrwkv.configuration_rwkv_world import RwkvConfig
+from ringrwkv.rwkv_tokenizer import TRIE_TOKENIZER
+from ringrwkv.modehf_world import RwkvForCausalLM
+
+
+if torch.cuda.is_available():
+    device = "cuda"
+else:
+    device = "cpu"
+
+#放在本地工程根目录文件夹
+
+
+model = RwkvForCausalLM.from_pretrained("RWKV-4-World-7B")
+tokenizer = TRIE_TOKENIZER('./ringrwkv/rwkv_vocab_v20230424.txt')
+
+
+#model= PeftModel.from_pretrained(model, "./lora-out")
+model = model.to(device)
+
+
+def evaluate(
+    instruction,
+    temperature=1,
+    top_p=0.7,
+    top_k = 0.1,
+    penalty_alpha = 0.1,
+    max_new_tokens=128,
+):
+    
+    prompt = f'Question: {instruction.strip()}\n\nAnswer:'
+    input_ids = tokenizer.encode(prompt)
+    input_ids = torch.tensor(input_ids).unsqueeze(0)
+    #out = model.generate(input_ids=input_ids.to(device),max_new_tokens=40)
+    out = model.generate(input_ids=input_ids.to(device),temperature=temperature,top_p=top_p,top_k=top_k,penalty_alpha=penalty_alpha,max_new_tokens=max_new_tokens)
+    outlist = out[0].tolist()
+    for i  in outlist:
+        if i==0:
+            outlist.remove(i)
+    answer = tokenizer.decode(outlist)
+    return answer.strip()
+    #return answer.split("### Response:")[1].strip()
+
+
+gr.Interface(
+    fn=evaluate,#接口函数
+    inputs=[
+        gr.components.Textbox(
+            lines=2, label="Instruction", placeholder="Tell me about alpacas."
+        ),
+        gr.components.Slider(minimum=0, maximum=2, value=1, label="Temperature"),
+        gr.components.Slider(minimum=0, maximum=1, value=0.7, label="Top p"),
+        gr.components.Slider(minimum=0, maximum=1, step=1, value=0.1, label="top_k"),
+        gr.components.Slider(minimum=0, maximum=1, step=1, value=0.1, label="penalty_alpha"),
+        gr.components.Slider(
+            minimum=1, maximum=2000, step=1, value=128, label="Max tokens"
+        ),
+    ],
+    outputs=[
+        gr.inputs.Textbox(
+            lines=5,
+            label="Output",
+        )
+    ],
+    title="RWKV-World-Alpaca",
+    description="RWKV,Easy In HF.",
+).launch()

hello_hf.py

@@ -0,0 +1,51 @@
+import torch
+from ringrwkv.configuration_rwkv_world import RwkvConfig
+from ringrwkv.rwkv_tokenizer import TRIE_TOKENIZER
+from ringrwkv.modehf_world import RwkvForCausalLM
+
+
+model = RwkvForCausalLM.from_pretrained("RWKV-4-World-7B")
+tokenizer = TRIE_TOKENIZER('./ringrwkv/rwkv_vocab_v20230424.txt')
+     
+text = "你叫什么名字？"
+
+question = f'Question: {text.strip()}\n\nAnswer:'
+
+input_ids = tokenizer.encode(question)
+#print(tokenizer.decode(input_ids))
+input_ids = torch.tensor(input_ids).unsqueeze(0)
+
+out = model.generate(input_ids,max_new_tokens=40)
+
+#print(out[0])
+
+outlist = out[0].tolist()
+
+for i  in outlist:
+    if i==0:
+        outlist.remove(i)
+
+#print(outlist)
+answer = tokenizer.decode(outlist)
+
+# answer = tokenizer.decode([10464, 11685, 19126, 12605, 11021, 10399, 12176, 10464, 16533, 10722,
+#          10250, 10349, 17728, 18025, 10080, 16738, 17728, 10464, 17879, 16503])
+# answer = tokenizer.decode([53648,    59,    33, 10464, 11017, 10373, 10303, 11043, 11860, 19156,
+#           261, 40301,    59,    33, 12605, 13091, 10250, 10283, 10370, 12137,
+#         13133, 15752, 16728, 16537, 13499, 11496, 19137, 13734, 13191, 11043,
+#         11860, 10080])
+print(answer)
+
+#print(input_ids.shape)
+#rwkvoutput = model.forward(input_ids=input_ids,labels=input_ids) #loss,logits,state,hidden_states,attentions
+# print("loss:")
+# print(rwkvoutput.loss)
+# print("logits:")
+# print(rwkvoutput.logits)
+# print("state:")
+# print(rwkvoutput.state)
+#print("last_hidden_state:")
+# print(rwkvoutput.last_hidden_state)
+# print("attentions:")
+# print(rwkvoutput.attentions)
+

ringrwkv.rar

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3d0f6c7c7e365ad46b5969eeec655ce733514ca06e1485c6453766156f456032
+size 261848