Upload 40 files

.gitattributes

@@ -34,3 +34,6 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
 *.axmodel filter=lfs diff=lfs merge=lfs -text
+main_ax650 filter=lfs diff=lfs merge=lfs -text
+main_axcl_aarch64 filter=lfs diff=lfs merge=lfs -text
+main_axcl_x86 filter=lfs diff=lfs merge=lfs -text

main_ax650

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:870254b83b53d8f1909a03d2e73b0049fd5541e848579fc530888d07c28a48a2
+size 989368

main_axcl_aarch64

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:03012513d42b834360fac6e61aa8b7387deb4f5ecd9f9c86d0b208e608b8c991
+size 1724632

main_axcl_x86

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5f872d9f2238ec7db7ce3a075d52a673bac7b5a7e18455a81852b80223037ae3
+size 1839824

minicpm4-0.5b-int8-ctx-ax650/MiniCPMForCausalLM_p128_l0_together.axmodel

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a464bc5480ab466fb81df78b7f38fc1d2afd3c07e6b8ed9e43e8a4661d265b1d
+size 17484788

minicpm4-0.5b-int8-ctx-ax650/MiniCPMForCausalLM_p128_l10_together.axmodel

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fa4feb9e333fccafbed59942743b0815089ecfef2efa83b8e6f2b1d53ecf855e
+size 17473812

minicpm4-0.5b-int8-ctx-ax650/MiniCPMForCausalLM_p128_l11_together.axmodel

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4b6592b72f251590db011e9923d6f51e64dce5eb25a937111d957371ee40587c
+size 17473812

minicpm4-0.5b-int8-ctx-ax650/MiniCPMForCausalLM_p128_l12_together.axmodel

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:eff9f8748fa09a9228e718160008d33f8e2c4ed335bc978c41471f09796843c5
+size 17473812

minicpm4-0.5b-int8-ctx-ax650/MiniCPMForCausalLM_p128_l13_together.axmodel

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dee4386ae92bdfeaa83590b2a5720598c944b75ea0eca296a5c1cd77f9328b76
+size 17473812

minicpm4-0.5b-int8-ctx-ax650/MiniCPMForCausalLM_p128_l14_together.axmodel

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f4ac5661daec71577bb0bd864d59f8bc63e076d6a3ee647b3cb7b3d2b87ed11d
+size 17473812

minicpm4-0.5b-int8-ctx-ax650/MiniCPMForCausalLM_p128_l15_together.axmodel

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fda43fc22089a66f2e7b91ce1dad1a587ab8c33057755aa7f34f977839b116b1
+size 17473812

minicpm4-0.5b-int8-ctx-ax650/MiniCPMForCausalLM_p128_l16_together.axmodel

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f55ee28760e580a88a5e62dfabdf1aaf8faf4205f7fb8434747cff1c0d6f3013
+size 17473812

minicpm4-0.5b-int8-ctx-ax650/MiniCPMForCausalLM_p128_l17_together.axmodel

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d2267345dc7101f1bb28ed4bc95d92b86de77b59a862e0bad0011ef27f4e23c0
+size 17473812

minicpm4-0.5b-int8-ctx-ax650/MiniCPMForCausalLM_p128_l18_together.axmodel

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5d6844de1eda71f56656588e76ab985520462b7b124172bdaaea426bf5f8ffbb
+size 17473812

minicpm4-0.5b-int8-ctx-ax650/MiniCPMForCausalLM_p128_l19_together.axmodel

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2f29a2a84421b83d722b3c0926ce4627f7328b80dfe4720f3baf487651cad02f
+size 17473812

minicpm4-0.5b-int8-ctx-ax650/MiniCPMForCausalLM_p128_l1_together.axmodel

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:189188e9b5a332e464b78a59faebde1707cff564836980e46812c03af8f17f80
+size 17473812

minicpm4-0.5b-int8-ctx-ax650/MiniCPMForCausalLM_p128_l20_together.axmodel

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bd589426ec6c7831e7a710038762118b7933951faa8c6edcf9a7a8f755d29b9c
+size 17473812

minicpm4-0.5b-int8-ctx-ax650/MiniCPMForCausalLM_p128_l21_together.axmodel

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ee832073759c4630cd3bfee9fb25dcb936c9be0ade8f8269630b2f89f863ad7e
+size 17473812

minicpm4-0.5b-int8-ctx-ax650/MiniCPMForCausalLM_p128_l22_together.axmodel

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9003410840fbf31810f13c54b0fb925decf202959b16da4e894eec8c57507fc5
+size 17473812

minicpm4-0.5b-int8-ctx-ax650/MiniCPMForCausalLM_p128_l23_together.axmodel

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:80a35d4f1a12d88b63c34202f321743f689290ac27885356c26bffdc960a7338
+size 17473812

minicpm4-0.5b-int8-ctx-ax650/MiniCPMForCausalLM_p128_l2_together.axmodel

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ecf6012ab296f5e4ea8a2d67a867d2c2a00cdcedeaf81df7e78651d3135325da
+size 17473812

minicpm4-0.5b-int8-ctx-ax650/MiniCPMForCausalLM_p128_l3_together.axmodel

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bac813c3ba1f660c2deb9f18a350fb7e8bf41219cb29091c67ee8805f51cb921
+size 17473812

minicpm4-0.5b-int8-ctx-ax650/MiniCPMForCausalLM_p128_l4_together.axmodel

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:08141250430f708078aeb76263eb253288e001483008b8e85cef2a5172eb2606
+size 17473812

minicpm4-0.5b-int8-ctx-ax650/MiniCPMForCausalLM_p128_l5_together.axmodel

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:40eeab8f4971717ce7570a549701d93bae6c8c70458dcba3e54eb59cc4676984
+size 17473812

minicpm4-0.5b-int8-ctx-ax650/MiniCPMForCausalLM_p128_l6_together.axmodel

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2f3c4a4e7040be172f266f521b990732dc3c45fd93750b7fd4be464e0b02153c
+size 17473812

minicpm4-0.5b-int8-ctx-ax650/MiniCPMForCausalLM_p128_l7_together.axmodel

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c31a22f0a4639df13cb7368f2f654a19d62b1693da75d24b352d4b81f247b2eb
+size 17473812

minicpm4-0.5b-int8-ctx-ax650/MiniCPMForCausalLM_p128_l8_together.axmodel

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b75ad009315ff06f5dd99eb3e24fc4ac4b6ce3686cb15614ccace73140c30c1b
+size 17473812

minicpm4-0.5b-int8-ctx-ax650/MiniCPMForCausalLM_p128_l9_together.axmodel

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:05a679d5a05cdaa0744711d4094a27327221ca99d2494dca697469a85a79b86a
+size 17473812

minicpm4-0.5b-int8-ctx-ax650/MiniCPMForCausalLM_post.axmodel

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1a98a8caef2c9769674a7390b1506695c38c25b970c551dc64369373274f8d89
+size 82179198

minicpm4-0.5b-int8-ctx-ax650/model.embed_tokens.weight.bfloat16.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:89e99a3acaab0bc7fe030abef0fa2c41e6373afbf40a79c84be6cb620ea5495d
+size 150421504

minicpm4_tokenizer/added_tokens.json

@@ -0,0 +1,10 @@
+{
+  "<|execute_end|>": 73444,
+  "<|execute_start|>": 73443,
+  "<|fim_middle|>": 73446,
+  "<|fim_prefix|>": 73445,
+  "<|fim_suffix|>": 73447,
+  "<|im_end|>": 73440,
+  "<|im_start|>": 73441,
+  "<|tool_call|>": 73442
+}

minicpm4_tokenizer/config.json

@@ -0,0 +1,37 @@
+{
+    "_name_or_path": "openbmb/MiniCPM4-0.5B",
+    "architectures": [
+        "MiniCPMForCausalLM"
+    ],
+    "auto_map": {
+        "AutoConfig": "configuration_minicpm.MiniCPMConfig",
+        "AutoModel": "modeling_minicpm.MiniCPMModel",
+        "AutoModelForCausalLM": "modeling_minicpm.MiniCPMForCausalLM",
+        "AutoModelForSeq2SeqLM": "modeling_minicpm.MiniCPMForCausalLM",
+        "AutoModelForSequenceClassification": "modeling_minicpm.MiniCPMForSequenceClassification"
+    },
+    "bos_token_id": 1,
+    "eos_token_id": [2, 73440],
+    "hidden_act": "silu",
+    "hidden_size": 1024,
+    "initializer_range": 0.1,
+    "intermediate_size": 4096,
+    "max_position_embeddings": 32768,
+    "num_attention_heads": 16,
+    "num_hidden_layers": 24,
+    "num_key_value_heads": 2,
+    "rms_norm_eps": 1e-05,
+    "rope_scaling": {
+        "rope_type": "longrope", 
+        "long_factor": [1.0004360675811768, 1.0668443441390991, 1.1631425619125366, 1.3025742769241333, 1.5040205717086792, 1.7941505908966064, 2.2101221084594727, 2.802666664123535, 3.6389970779418945, 4.804192543029785, 6.39855432510376, 8.527148246765137, 11.277542114257812, 14.684998512268066, 18.69317054748535, 23.13019371032715, 27.72362518310547, 32.1606559753418, 36.168827056884766, 39.57627868652344, 42.32667541503906, 44.45526885986328, 46.04962921142578, 47.21482849121094, 48.05115509033203, 48.64370346069336, 49.05967712402344, 49.34980392456055, 49.551246643066406, 49.69068145751953, 49.78697967529297, 49.85338592529297],
+        "short_factor": [1.0004360675811768, 1.0668443441390991, 1.1631425619125366, 1.3025742769241333, 1.5040205717086792, 1.7941505908966064, 2.2101221084594727, 2.802666664123535, 3.6389970779418945, 4.804192543029785, 6.39855432510376, 8.527148246765137, 11.277542114257812, 14.684998512268066, 18.69317054748535, 23.13019371032715, 27.72362518310547, 32.1606559753418, 36.168827056884766, 39.57627868652344, 42.32667541503906, 44.45526885986328, 46.04962921142578, 47.21482849121094, 48.05115509033203, 48.64370346069336, 49.05967712402344, 49.34980392456055, 49.551246643066406, 49.69068145751953, 49.78697967529297, 49.85338592529297],
+        "original_max_position_embeddings": 32768
+    },    
+    "torch_dtype": "bfloat16",
+    "transformers_version": "4.46.3",
+    "use_cache": true,
+    "vocab_size": 73448,
+    "scale_emb": 12,
+    "dim_model_base": 256,
+    "scale_depth": 1.4
+}

minicpm4_tokenizer/generation_config.json

@@ -0,0 +1,12 @@
+{
+  "bos_token_id": 1,
+  "do_sample": true,
+  "eos_token_id": [
+    2,
+    73440
+  ],
+  "pad_token_id": 2,
+  "temperature": 0.8,
+  "top_p": 0.8,
+  "transformers_version": "4.46.1"
+}

minicpm4_tokenizer/special_tokens_map.json

@@ -0,0 +1,33 @@
+{
+  "additional_special_tokens": [
+    "<|im_end|>",
+    "<|im_start|>",
+    "<|tool_call|>",
+    "<|execute_start|>",
+    "<|execute_end|>",
+    "<|fim_prefix|>",
+    "<|fim_middle|>",
+    "<|fim_suffix|>"
+  ],
+  "bos_token": {
+    "content": "<s>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "<|im_end|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "unk_token": {
+    "content": "<unk>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

minicpm4_tokenizer/tokenizer.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bb74d51116831c3bf65db812c553f94ab0c88dcf97a5bbb37e3504f6d359c530
+size 1181204

minicpm4_tokenizer/tokenizer_config.json

@@ -0,0 +1,117 @@
+{
+  "add_bos_token": true,
+  "add_eos_token": false,
+  "add_prefix_space": null,
+  "added_tokens_decoder": {
+    "0": {
+      "content": "<unk>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "1": {
+      "content": "<s>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "2": {
+      "content": "</s>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "73440": {
+      "content": "<|im_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "73441": {
+      "content": "<|im_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "73442": {
+      "content": "<|tool_call|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "73443": {
+      "content": "<|execute_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "73444": {
+      "content": "<|execute_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "73445": {
+      "content": "<|fim_prefix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "73446": {
+      "content": "<|fim_middle|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "73447": {
+      "content": "<|fim_suffix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "additional_special_tokens": [
+    "<|im_end|>",
+    "<|im_start|>",
+    "<|tool_call|>",
+    "<|execute_start|>",
+    "<|execute_end|>",
+    "<|fim_prefix|>",
+    "<|fim_middle|>",
+    "<|fim_suffix|>"
+  ],
+  "bos_token": "<s>",
+  "chat_template": "{% for message in messages %}{{'<|im_start|>' + message['role'] + '\n' + message['content'] + '<|im_end|>' + '\n'}}{% endfor %}{% if add_generation_prompt %}{{ '<|im_start|>assistant\n' }}{% endif %}",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|im_end|>",
+  "legacy": true,
+  "model_max_length": 1000000000000000019884624838656,
+  "pad_token": null,
+  "sp_model_kwargs": {},
+  "spaces_between_special_tokens": false,
+  "tokenizer_class": "LlamaTokenizer",
+  "unk_token": "<unk>",
+  "use_default_system_prompt": false
+}

minicpm4_tokenizer_uid.py

@@ -0,0 +1,199 @@
+from transformers import AutoTokenizer, PreTrainedTokenizerFast
+from http.server import HTTPServer, BaseHTTPRequestHandler
+import json
+import argparse
+import uuid
+
+# 全局字典：存储 uid 到 Tokenizer_Http 实例的映射
+tokenizers = {}
+
+class Tokenizer_Http():
+    def __init__(self):
+        model_id = "minicpm4_tokenizer"
+        self.tokenizer = AutoTokenizer.from_pretrained(model_id)
+        self.messages = [
+            {"role": "system", "content": "You are MiniCPM4, created by ModelBest. You are a helpful assistant."},
+        ]
+        self.token_ids = []
+        
+        self.token_ids_cache = []
+
+    def encode(self, prompt, last_reply=None):
+        if last_reply is not None:
+            self.messages.append({"role": "assistant", "content": last_reply})
+            text = self.tokenizer.apply_chat_template(
+                self.messages,
+                tokenize=False,
+                add_generation_prompt=True
+            )
+            # print("生成的文本:\n============\n", text, "============\n")
+            self.token_ids = self.tokenizer.encode(text)[:-3]
+        self.messages.append({"role": "user", "content": prompt})
+        
+        text = self.tokenizer.apply_chat_template(
+            self.messages,
+            tokenize=False,
+            add_generation_prompt=True
+        )
+        print("生成的文本:\n============\n", text, "============\n")
+        token_ids = self.tokenizer.encode(text)
+        # 找出新增部分
+        diff = token_ids[len(self.token_ids):]
+        self.token_ids = token_ids
+        print(self.decode(diff))
+        return token_ids, diff
+
+    def decode(self, token_ids):
+        self.token_ids_cache += token_ids
+        text = self.tokenizer.decode(self.token_ids_cache)
+        if "\ufffd" in text:
+            print("text 中包含非法字符")
+            return ""
+        else:
+            self.token_ids_cache.clear()
+            return text
+        
+
+    @property
+    def bos_id(self):
+        return self.tokenizer.bos_token_id
+
+    @property
+    def eos_id(self):
+        return self.tokenizer.eos_token_id
+    
+    @property
+    def bos_token(self):
+        return self.tokenizer.bos_token
+
+    @property
+    def eos_token(self):
+        return self.tokenizer.eos_token
+    
+    def reset(self, system_prompt="You are MiniCPM4, created by ModelBest. You are a helpful assistant."):
+        self.messages = [
+            {"role": "system", "content": system_prompt},
+        ]
+        text = self.tokenizer.apply_chat_template(
+            self.messages,
+            tokenize=False,
+            add_generation_prompt=True
+        )
+        token_ids = self.tokenizer.encode(text)[:-3]
+        self.token_ids = token_ids
+        print(self.decode(token_ids))
+        return token_ids
+        
+
+class Request(BaseHTTPRequestHandler):
+    timeout = 5
+    server_version = 'Apache'
+
+    def do_GET(self):
+        print("GET 请求路径:", self.path)
+        self.send_response(200)
+        self.send_header("Content-Type", "application/json")
+        self.end_headers()
+
+        # 新增接口：获取 uid
+        if '/get_uid' in self.path:
+            new_uid = str(uuid.uuid4())
+            print("新 uid:", new_uid)
+            # 为该 uid 创建一个新的 Tokenizer_Http 实例
+            tokenizers[new_uid] = Tokenizer_Http()
+            msg = json.dumps({'uid': new_uid})
+        elif '/bos_id' in self.path:
+            # 获取 uid 参数（例如 ?uid=xxx）
+            uid = self.get_query_param("uid")
+            instance: Tokenizer_Http = tokenizers.get(uid)
+            if instance is None:
+                msg = json.dumps({'error': 'Invalid uid'})
+            else:
+                bos_id = instance.bos_id
+                msg = json.dumps({'bos_id': bos_id if bos_id is not None else -1})
+        elif '/eos_id' in self.path:
+            uid = self.get_query_param("uid")
+            instance: Tokenizer_Http = tokenizers.get(uid)
+            if instance is None:
+                msg = json.dumps({'error': 'Invalid uid'})
+            else:
+                eos_id = instance.eos_id
+                msg = json.dumps({'eos_id': eos_id if eos_id is not None else -1})
+        else:
+            msg = json.dumps({'error': 'Invalid GET endpoint'})
+
+        print("响应消息:", msg)
+        self.wfile.write(msg.encode())
+
+    def do_POST(self):
+        content_length = int(self.headers.get('content-length', 0))
+        data = self.rfile.read(content_length).decode()
+        print("POST 请求路径:", self.path)
+        print("接收到的数据:", data)
+        req = json.loads(data)
+
+        self.send_response(200)
+        self.send_header("Content-Type", "application/json")
+        self.end_headers()
+
+        if '/encode' in self.path:
+            # 请求数据中必须包含 uid, text, 和可选的 last_reply
+            uid = req.get('uid')
+            prompt = req.get('text')
+            last_reply = req.get('last_reply')
+            instance: Tokenizer_Http = tokenizers.get(uid)
+            if instance is None:
+                msg = json.dumps({'error': 'Invalid uid'})
+            else:
+                token_ids, diff = instance.encode(prompt, last_reply)
+                msg = json.dumps({'token_ids': token_ids, 'diff': diff})
+        elif '/decode' in self.path:
+            uid = req.get('uid')
+            token_ids = req.get('token_ids')
+            instance: Tokenizer_Http = tokenizers.get(uid)
+            if instance is None:
+                msg = json.dumps({'error': 'Invalid uid'})
+            else:
+                text = instance.decode(token_ids)
+                msg = json.dumps({'text': text})
+        elif '/reset' in self.path:
+            uid = req.get("uid")
+            system_prompt = req.get("system_prompt")
+            instance: Tokenizer_Http = tokenizers.get(uid)
+            if instance is None:
+                msg = json.dumps({'error': 'Invalid uid'})
+            else:
+                if system_prompt is not None:
+                    print("system_prompt:", system_prompt)
+                    token_ids = instance.reset(system_prompt)
+                    msg = json.dumps({'token_ids': token_ids})
+                else:
+                    token_ids = instance.reset()
+                    msg = json.dumps({'token_ids': token_ids})
+        else:
+            msg = json.dumps({'error': 'Invalid POST endpoint'})
+
+        print("响应消息:", msg)
+        self.wfile.write(msg.encode())
+
+    def get_query_param(self, key):
+        """
+        辅助函数：从 GET 请求的 URL 中获取查询参数的值
+        例如：/bos_id?uid=xxx
+        """
+        from urllib.parse import urlparse, parse_qs
+        query = urlparse(self.path).query
+        params = parse_qs(query)
+        values = params.get(key)
+        return values[0] if values else None
+
+if __name__ == "__main__":
+    parser = argparse.ArgumentParser()
+    parser.add_argument('--host', type=str, default='0.0.0.0')
+    parser.add_argument('--port', type=int, default=12345)
+    args = parser.parse_args()
+
+    host = (args.host, args.port)
+    print('Server running at http://%s:%s' % host)
+    server = HTTPServer(host, Request)
+    server.serve_forever()

post_config.json

@@ -0,0 +1,14 @@
+{
+    "enable_temperature" : false,
+    "temperature" : 0.9,
+    
+    "enable_repetition_penalty" : false,
+    "repetition_penalty" : 1.2,
+    "penalty_window" : 20,
+
+    "enable_top_p_sampling" : false,
+    "top_p" : 0.8,
+
+    "enable_top_k_sampling" : true,
+    "top_k" : 1
+}

run_minicpm4_0.5b_int8_ctx_ax650.sh

@@ -0,0 +1,14 @@
+./main_ax650 \
+--system_prompt "You are MiniCPM4, created by ModelBest. You are a helpful assistant." \
+--template_filename_axmodel "minicpm4-0.5b-int8-ctx-ax650/MiniCPMForCausalLM_p128_l%d_together.axmodel" \
+--axmodel_num 24 \
+--tokenizer_type 2 \
+--url_tokenizer_model "http://127.0.0.1:12345" \
+--filename_post_axmodel "minicpm4-0.5b-int8-ctx-ax650/MiniCPMForCausalLM_post.axmodel" \
+--filename_tokens_embed "minicpm4-0.5b-int8-ctx-ax650/model.embed_tokens.weight.bfloat16.bin" \
+--tokens_embed_num 73448 \
+--tokens_embed_size 1024 \
+--use_mmap_load_embed 0 \
+--live_print 1
+
+# --kvcache_path /home/axera/ax-llm/build/kvcache_yuanqi \